Nontuberculous Mycobacteria - ATS Journals

Nontuberculous Mycobacteria Defining Disease in a Prospective Cohort of South African Miners ELIZABETH L. CORBETT, LUCILLE BLUMBERG, GAVIN J. CHURCHYARD, NOMSA MOLOI, KATHY MALLORY, TIM CLAYTON, BRIAN G. WILLIAMS, RICHARD E. CHAISSON, RICHARD J. HAYES, and KEVIN M. DE COCK Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, United Kingdom; Department of Mycobacteriology, South African Institute of Medical Research, Department of Medicine, Ernest Oppenheimer Hospital, Welkom; Epidemiology Research Unit, Johannesburg, South Africa; and School of Medicine, Johns Hopkins University, Baltimore, Maryland

A gold mining work force was followed prospectively over 1 yr for sputum nontuberculous mycobacterial (NTM) isolates. NTM were isolated from 118 men, of whom 32 (27%) met the American Thoracic Society (ATS) case-definitions for pulmonary NTM disease (23 M. kansasii, seven M. scrofulaceum, one M. avium, and one M. abscessus). Determining isolate significance was difficult because most men had been started on presumptive TB treatment before isolate identification (70%). Histologic criteria were considered inappropriate for this high M. tuberculosis incidence population, particularly for patients who had stabilized on presumptive TB treatment. Among men not meeting casedefinitions, indicators of disease were significantly more prevalent for M. kansasii than for M. fortuitum, the local laboratory contaminant (ORs: 6.5 for cough, 7.2 for smear-positivity, 36.0 for radiologic changes, and 14.3 for presumptive TB treatment), suggesting underdiagnosis of M. kansasii disease. Of 53 men with definite or possible M. kansasii disease, 18 (34%) were HIV-positive. HIV-associated M. kansasii disease occurred at an early stage of immunosuppression (median CD4 count, 381 3 106/L) with a good outcome (83% cured after 12 mo of treatment). ATS case-definitions for NTM disease are difficult to apply in this population, and treatment decisions should be guided by the pathogenic potential of the isolate. Corbett EL, Blumberg L, Churchyard GJ, Moloi N, Mallory K, Clayton T, Williams BG, Chaisson RE, Hayes RJ, De Cock KM. Nontuberculous mycobacteria: defining disease in a prospective cohort of South African miners. AM J RESPIR CRIT CARE MED 1999;160:15–21.

The low incidence of pulmonary nontuberculous mycobacterial (NTM) isolates and disease in developed countries has made prospective studies concerning the significance of sputum NTM isolates difficult to conduct. In less developed countries, very little information exists about pulmonary NTM disease incidence or mycobacteriology, mainly because mycobacterial culture and identification are too expensive to be routine practice in resource-poor health systems. South African gold miners have a high incidence of M. tuberculosis disease, and they are provided with health care that includes mycobacterial culture and isolate identification in cases of suspected tuberculosis (TB). Previous studies have shown that isolates of the most frequent NTM species, M. kansasii and M. scrofulaceum, occur with high incidence (1) and are usually associated with positive sputum smears, new radiologic changes, and a significantly higher prevalence of NTM risk factors such as silicosis and posttuberculosis lung disease than either patients with TB or control patients (2). HIV is

also a significant risk factor for pulmonary NTM isolates in miners (1, 2). However, despite these indicators of pathogenicity, uncertainty about the clinical significance of NTM isolates in miners remains since these studies used retrospective data collected when standard management was to identify a single mycobacterial isolate per disease episode. Because single sputum NTM isolates can result from specimen contamination or transient infection, a firm diagnosis of NTM disease requires multiple isolates as well as evidence of disease (3). The results presented here are from a prospective cohort study of NTM isolates in miners, using the American Thoracic Society (ATS) case-definitions for diagnosis of NTM disease (3). The main aim was to investigate the clinical significance of different NTM isolates in HIV-positive and HIV-negative miners. The study also provided an opportunity to assess the feasibility of applying the ATS case-definitions prospectively in a population with a high incidence of M. tuberculosis disease.

METHODS (Received in original form December 9, 1998 and in revised form February 22, 1999) Dr. Corbett is the recipient of a Fellowship in Clinical Tropical Medicine from the Wellcome Trust. Correspondence and requests for reprints should be addressed to Dr. E. L. Corbett, Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, UK. Am J Respir Crit Care Med Vol 160. pp 15–21, 1999 Internet address: www.atsjournals.org

Study Population Miners working for a single mining company in the Free State province, South Africa are provided with free health care, including a centralized TB service based at a single hospital. Men with suspected pulmonary mycobacterial disease were investigated according to standard hospital policy with three sputum samples collected over 2 d and a chest radiograph. Men with positive sputum smears were started on

16

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

presumptive TB treatment, and those with negative sputum smears were treated with 1 wk of ampicillin and followed up with their sputum culture results.

Bacteriology Sputum samples were processed with a concentration and decontamination step and stained with auramine for fluorescent microscopy. Positive slides were confirmed with Ziehl-Neelsen (ZN) staining. All sputum samples were inoculated onto Lowenstein-Jensen (LJ) slopes and incubated for as long as 8 wk. An initial identification step for M. tuberculosis was carried out on LJ slopes with more than five colonies, using a calorimetric ribosomal RNA hybridization test (Accuprobe M. tuberculosis complex probe kit; Gen-Probe, San Diego, CA). LJ slopes with less than five colonies were sent directly to the South African Institute of Medical Research (SAIMR) mycobacteriology laboratory for biochemical species identification (4, 5).

Initial Assessment Patients with mycobacterial isolates that were negative for M. tuberculosis RNA hybridization or identified biochemically as being NTM species between December 1996 and November 1997 were recalled to the study clinic and assessed for clinical and bacteriologic evidence of pulmonary NTM disease, using the ATS case-definitions for NTM disease (summarized in the APPENDIX). Follow-up was continued until November 1998. Patients were questioned for symptoms, past mycobacterial disease, and occupational and smoking history, and were weighed. HIV tests were taken, with informed consent and pretest and post-test counselling from all patients. HIV infection was diagnosed if both screening (Enzyme-Test; Boehringer Mannheim Immunodiagnostics, Indianapolis, IN) and confirmatory ELISA (IMx system; Abbott Diagnostics, North Chicago, IL) were positive. Blood for CD41 T-lymphocyte measurement (CD4 count) was taken with consent in all patients with a positive HIV-test (FACScount; Becton Dickinson, Mountain View, CA). All men had two further sputum samples taken for microscopy and culture onto LJ slopes. Men who were not already receiving presumptive TB treatment had one further sputum specimen cultured into BACTEC medium. Men were followed monthly until identification of all available cultures. Treated patients receiving stable regimens were then seen every 3 mo. Untreated men with stable radiographs were seen every 1 to 2 mo for 6 mo and then every 3 mo for as long as 12 mo. Sputum for microscopy and LJ culture and chest radiographs were taken on each visit. Two sputum samples were taken for microscopy and LJ culture on completion of treatment. An additional sputum specimen for BACTEC culture was taken at each clinic visit from untreated men. Routine annual screening miniradiographs from at least 1 yr before the NTM isolate and presenting radiographs from the time of the positive culture were scored by two readers for the presence of preexisting NTM disease risk factors (silicosis and focal scarring suggestive of healed TB) and new radiologic changes associated with the isolate. Radiographs were read without reference to clinical or bacteriologic data. Radiologic response to treatment was assessed and recorded at each clinic attendance by the attending physician. An on-site high resolution spiral CT scanner and facilities for bronchoscopy were available, but they were not used except two patients where nonmycobacterial pathology was suspected.

Treatment Regimens and Outcomes Treatment regimens are summarized in Table 1. Presumptive TB regimens were used to treat patients with mixed M. tuberculosis and NTM isolates and also those with non-kansasii NTM isolates who did not meet the ATS case-definitions for NTM disease but were considered to have mycobacterial disease on clinical and radiologic grounds. Patients with M. kansasii isolates associated with new radiologic changes were treated for M. kansasii disease even if ATS case-definitions were not met. Prophylactic cotrimoxazole, 960 mg daily, was offered to all HIV-positive patients with a CD4 count below 200 3 106/L. Treatment was given as directly observed therapy. Treatment outcomes were defined as: cured (negative sputum culture within a month of stopping treatment); failed (positive sputum culture within a month of stopping treatment); died (died while receiving treatment);

VOL 160

1999

transferred out (left the mining industry while receiving treatment and not followed up at the study clinic).

Data Analysis Company payroll records were used to calculate the total number of employee-years worked for the company during the study period in order to estimate NTM disease incidence. STATA 5.0 (STATA Corp., Austin, TX) and EPI-INFO 6.02 (CDC, Atlanta, GA) software were used for data analysis. Differences in categorical variables between subgroups were expressed as odds ratios (OR) and tested for significance using exact 95% confidence intervals (95% CI). Student’s t test was used to test for differences between mean age and employment duration for patients with different NTM species. The Mann-Whitney two-sample test was used to test for group differences in median CD4 counts. The study was approved by the Ethics Committees of the London School of Hygiene and Tropical Medicine and the Ernest Oppenheimer Hospital.

RESULTS NTM Isolates and Patient Numbers

NTM were isolated from 118 patients during the study period of whom 40 (34%) were HIV-positive. A further 13 patients, who were not considered further, were investigated because of a provisional report of an NTM isolate, which was M. tuberculosis on final identification. The number of patients with different NTM species are shown in Table 2, together with the prevalence of HIV infection, smear positivity (any occasion), proportion meeting the ATS case-definitions, and proportion of potential relapsed and chronic cases. Thirty-two patients (27%) met the ATS case-definitions for pulmonary NTM disease. Sixty-nine (58%) were smear-positive. Nineteen (16%) had previously been treated for pulmonary disease associated with the same NTM species (potential relapses), including five (4%) with chronic culture-positive NTM disease and both patients with mixed M. kansasii and M. scrofulaceum isolates. Twenty-one (18%) had mixed mycobacterial species isolated, NTM and M. tuberculosis in 19 (16%) and M. kansasii and M. scrofulaceum in two (2%). All chronic cases and potential relapses had M. kansasii, M. scrofulaceum, or M. avium isolates, and all but one of

TABLE 1 TREATMENT REGIMENS

Presumptive TB† New case Retreatment case NTM disease M. kansasii i M. scrofulaceum

Intensive Phase*

Continuation Phase

RHZE‡ RHZES§

RHE RHE

None None

RHE RHE Clarith¶

Duration (mo) 6 8 12 15–18**

* First 2 mo of presumptive TB regimen. † Regimen used to treat patients with mixed M. tuberculosis and NTM isolates and also those with non-kansasii NTM isolates who did not meet the ATS case-definitions for NTM disease, but were considered to have mycobacterial disease on clinical and radiologic grounds. ‡ Drug abbreviations: R 5 rifampicin, H 5 isoniazid, Z 5 pyrazinamide, E 5 ethambutol, S 5 streptomycin. Fixed combination RHE tablets were used (150 mg of rifampicin, 75 mg of isoniazid, and 300 mg of ethambutol. Daily dose: 4 tablets if above 50 kg, 3 tablets if below 50 kg). § Streptomycin omitted if HIV-positive. i Regimen used for all patients with M. kansasii isolates and clinical and radiologic mycobacterial disease. ¶ Clarith 5 clarithromycin. ** Treatment was continued to 12 mo of culture-negativity.

17

Corbett, Blumberg, Churchyard, et al.: Nontuberculous Mycobacteria in Miners TABLE 2 PATIENTS’ HIV STATUS, SMEAR POSITIVITY, AND NUMBER OF PATIENTS MEETING THE ATS CASE-DEFINITIONS FOR NTM DISEASE ACCORDING TO MYCOBACTERIAL SPECIES

NTM Species

HIV Positive

Smear Positive

ATS Disease*

Patients (n)

n

%

n

%

n

56 14 17 10

19 2 7 4

34 14 41 40

41 10 3 6

73 71 18 60

23 7

19 2

7 1

37 50

9 0

47 0

118

40

34

69

58

§

M. kansasii M. scrofulaceum§ M. fortuitum§ Other NTM speciesi Mixed M. tuberculosis 1 NTM¶ Mixed NTM isolates Total

2

% 41 50 0 20 0** N/A††

32

27

Potential Relapse†

Chronic Disease‡

n

%

n

%

10 4

2 2

1

18 29 0 10

4 14 0 10

2 2

11 100

19

16

1

0 0 5

4

* Number of patients with isolate who met the ATS case-definitions for NTM disease by the end of their follow-up. † Previously treated for the same NTM species, but not necessarily meeting ATS case-definitions for NTM disease on either occasion. Noncontributory to current ATS case-definition assessment in all cases because of time interval of more than 12 mo. ‡ Culture-positive while receiving treatment for previously diagnosed NTM disease. § Not including patients with mixed mycobacterial species isolates. i 1 M. abscessus, 3 M. avium, 2 M. gordonae, 1 M. szulgai, 1 M. malmoense, 1 M. flavescens, 1 M. terrae. ¶ 11 M. kansasii, 2 M. fortuitum, 2 M. scrofulaceum, 2 M. avium, 1 M. flavescens, 1 M. terrae. ** Clinical significance of NTM isolates coisolated with M. tuberculosis cannot be assessed until treated for M. tuberculosis under the ATS case-definitions. †† Mixed NTM isolates are not covered by ATS case-definitions. Both mixed M. kansasii and M. scrofulaceum.

the patients meeting the ATS case-definitions had these same species, the exception having M. abscessus disease. HIV Prevalence and CD4 Counts

HIV prevalence did not differ significantly between NTM groups, and there was no significant difference in the proportion of patients meeting the ATS case-definitions for NTM disease by HIV status (35% of HIV-negative and 27% of HIV-positive patients with single species isolates). Eight of the nine HIV-positive patients meeting ATS case-definitions had M. kansasii disease, and one had chronic culture-positive pulmonary M. avium disease. The median CD4 count for all HIV-positive patients with an NTM isolate was 322 3 106/L, with a range of 18 to 842 3 106/L. Only six (15%) of HIV-positive patients had a CD4 count below 100 3 106/L. Incidence

The number of employee-years worked for the company during the study period was 56,745. The overall incidence of NTM disease meeting ATS case-definitions (definite disease), excluding chronic cases, was 47.6 per 100,000 employee-years. The incidence rates for definite M. kansasii M. scrofulaceum disease were 37.0 and 8.8 per 100,000 employee-years, respectively. Radiologic and Bacteriologic Findings on First Clinic Recall

Radiologic and bacteriologic findings as shown in Table 3, with patients divided into groups according to initial NTM isolate and radiologic and treatment status when first recalled for NTM assessment. Patients were reassessed using ATS casedefinitions at each clinic visit, and their final assessment is the one used for Tables 2 and 3. Sixteen patients (14% of total) were untreated and had stable radiographs. Three (19% of subgroup) were again culture-positive (one M. kansasii, one mixed M. kansasii and M. scrofulaceum with multiple isolates of both species, and one M. tuberculosis after an initial isolate of M. scrofulaceum). Each developed new radiologic disease soon after assessment. Twelve patients (10% of total) were untreated but had symptoms and new radiologic changes that had progressed be-

tween first presentation and NTM assessment. Ten (83% of subgroup) were again culture-positive (four M. kansasii, two mixed M. kansasii and M. tuberculosis, two M. tuberculosis, and two M. scrofulaceum). Eighty-three patients (70% of total) had been started on presumptive TB before the culture results were available. Treatment had been started 1 to 22 wk before NTM assessment (median, 8 wk). Nine patients (11%) had less than three sputum samples taken for culture before treatment (six with a single smear-positive sample, and three with two samples). Only two patients, with M. scrofulaceum and M. abscessus, respectively, remained culture-positive. Multiple mycobacterial isolates were available from sputum collected previously in 32 patients, of whom 19 had multiple isolates of the same NTM species and 13 had mixed isolates of NTM species and M. tuberculosis. Radiologic Response to Treatment

There was a radiologic response in 47 (89%) of the 53 patients treated for M. kansasii isolates, six (55%) of the 11 patients with M. scrofulaceum isolates who received presumptive TB or NTM disease treatment, and six (75%) of the eight patients with M. fortuitum isolates who received presumptive TB treatment. No patient with these three NTM species had a radiologic deterioration during treatment, apart from one man with M. kansasii who developed a new pleural effusion, found to be due to metastatic adenocarcinoma, after having an initial radiologic response to treatment. Comparison between Patients with Different NTM Isolates

None of the patients with M. fortuitum had more than one isolate, and so none met the ATS case-definitions. There was suggestive evidence that these isolates represented laboratory contaminants, in that they occurred in clusters, with uniformly low colony counts from specimens with sequential or closely grouped laboratory numbers that temporally coincided with building work in close proximity to the laboratory. There was no clustering of specimen numbers for isolates of other NTM

18


VOL 160

1999

TABLE 3 INITIAL RADIOLOGIC ASSESSMENT, TREATMENT STATUS, AND SUBSEQUENT SPUTUM CULTURE RESULTS AFTER IDENTIFICATION OF A SPUTUM MYCOBACTERIAL ISOLATE AS AN NTM Initial NTM Assessment and Mycobacterial Isolate(s)

Further Culture Results

Patients (n)

ATS Disease (% subgroup)

Untreated: stable radiograph M. kansasii M. kansasii M. scrofulaceum M. scrofulaceum M. scrofulaceum M. fortuitum M. avium M. triviale

M. kansasii Negative M. kansasii 1 M. scrofulaceum M. tuberculosis Negative Negative Negative Negative

16 1 1 1 1 1 9 1 1

1 (6%) 1* 0 N/A*,† 0* 0 0 0 0

Untreated: progressive radiologic disease M. kansasii M. kansasii M. kansasii M. scrofulaceum M. fortuitum M. fortuitum M. triviale M. kansasii 1 M. scrofulaceum

M. kansasii Negative M. kansasii 1 M. tuberculosis M. scrofulaceum Negative M. tuberculosis M. tuberculosis M. scrofulaceum

12 4 1 2 1 1 1 1 1

4 (33%) 3 0 0 1 0 0 0 N/A†

Presumptive treatment already started M. kansasii M. scrofulaceum M. scrofulaceum M. abscessus M. fortuitum Other species Mixed NTM 1 M. tuberculosis

Negative M. scrofulaceum Negative M. abscessus Negative Negative Negative

83 47 1 9 1 7 5 13

22 (27%) 17 1 1 1 2‡ 1 0 0 0

Chronic NTM disease M. kansasii M. scrofulaceum M. avium

M. kansasii M. scrofulaceum M. avium

5 2 2 1

5 (100%) 2 2 1

* New radiologic changes within 2 mo of follow-up. † Multiple mixed isolates: not covered by ATS case-definitions. ‡ One who met ATS case-definitions initially and two who did so on early relapse after completing presumptive TB treatment.

species. Specimens of laboratory water and culture of noninoculated LJ medium did not yield any NTM species. There was a significant difference in the mean ages and duration of employment as a miner between patients with M. fortuitum isolates and M. kansasii (mean age, 41 and 47 yr, respectively, p 5 0.008; mean employment duration, 18 and 26 yr, respectively, p 5 0.01). Patients with M. scrofulaceum were also older (mean, 47 yr; p 5 0.047) and had worked for longer (mean, 26 yr; p 5 0.005) than patients with M. fortuitum. The clinical and radiologic features suggestive of mycobacterial disease in the 33 men with M. kansasii isolates and the 17 men with M. fortuitum isolates who did not meet the ATS case-definitions for pulmonary NTM disease for reasons other than coisolation of M. tuberculosis are shown in Table 4. These patients failed to meet ATS case-definitions because of: no evidence of disease plus a single isolate (one with M. kansasii and nine with M. fortuitum); new radiologic changes plus a single isolate only (30 with M. kansasii and eight with M. fortuitum); or new radiologic changes plus negative sputum smears with only two isolates (two patients with M. kansasii). Six of the patients with M. kansasii and one of those with M. fortuitum had less than three sputum cultures taken before starting treatment, and 10 of the patients with M. kansasii and two of those with M. fortuitum had one or more cultures that could not be assessed because of bacterial or fungal overgrowth (seven patients: six with M. kansasii and one with M. fortuitum) or mixed mycobacterial and bacterial colonies where the mycobacterial isolate could not be identified be-

cause of failure to grow when subcultured (five patients: four with M. kansasii and one with M. fortuitum). In total, 15 (45%) of those with M. kansasii and three (18%) of those with M. fortuitum were assessed on the results of less than three sputum cultures because of one or the other of these factors, and eight (24%) of the patients with M. kansasii were assessed

TABLE 4 COMPARISON OF CLINICAL AND RADIOLOGIC FEATURES SUGGESTIVE OF MYCOBACTERIAL DISEASE IN PATIENTS WITH M. kansasii OR M. fortuitum ISOLATES WHO DID NOT MEET THE ATS CASE DEFINITIONS FOR NTM DISEASE* M. kansasii (n 5 33)

M. fortuitum (n 5 17)

Feature

n

%

n

%

Odds Ratio

95% CI

Night sweats Worsening cough Daily sputum Hemoptysis Recent weight loss . 3 kg Smear-positive New chest radiograph changes New cavity TB treatment started before culture results

22 22 23 9 17 20

67 67 70 27 52 61

9 4 9 3 3 3

53 24 53 18 18 18

1.8 6.5 2.0 1.8 5.0 7.2

0.45–6.9 1.48–32.7 0.51–8.0 0.35–11.6 1.06–31.0 1.51–44.7

32 22

97 67

8 3

47 18

36.0 9.3

3.70–1,625 1.93–58.4

30

91

7

41

14.3

2.58–95.6

* Excluding men with M. tuberculosis coisolates.

Corbett, Blumberg, Churchyard, et al.: Nontuberculous Mycobacteria in Miners

on the results of a single culture. Indicators of mycobacterial disease were more common in the M. kansasii group than in the M. fortuitum group. Significant differences were for worsening cough (OR, 6.5); recent weight loss (OR, 5.0); smearpositivity (OR, 7.2); new radiologic changes (OR, 36.0); new cavitation (OR, 9.3). A decision to treat for presumptive TB before culture results were available had been taken in a significantly higher proportion of patients with M. kansasii than those with M. fortuitum (OR, 14.3; 95% CI, 2.6 to 95.6). Apart from smear-positivity and number of isolates, there were no significant differences between the prevalence of indicators of mycobacterial disease in patients with M. kansasii who did meet the ATS case-definitions and those who did not (data not shown). Indicators of mycobacterial disease were also common among the nine patients with M. scrofulaceum isolates who did not initially meet the ATS case-definitions for NTM disease, although further analysis was limited by small numbers. New radiographic changes were present in eight, with cavitation in six, six were smear-positive, and eight had been started on presumptive TB treatment before culture results were available. Comparison of HIV-positive and HIV-negative Patients with M. kansasii

The clinical and radiologic features of the 18 HIV-positive and 35 HIV-negative patients with definite or possible (treated patients with new radiologic changes who did not meet ATS case-definitions and did not have M. tuberculosis coisolates) M. kansasii disease did not differ significantly. Twelve (67%) HIV-positive and 27 (77%) HIV-negative patients were smear-positive (OR, 0.74; 95% CI, 0.19 to 3.10). New cavitation was present in ten (56%) of the HIV-positive and in 28 (80%) of the HIV-negative patients (OR, 0.40; 95% CI, 0.10 to 1.58) and new pleural thickening was present in four (22%) HIV-positive and in 14 (40%) HIV-negative patients (OR, 0.47; 95% CI, 0.09 to 1.94). The median CD4 count in HIV-positive patients with definite or possible M. kansasii disease was 381 3 106/L (range, 52 to 842 3 106/L), with only one patient having a CD4 count below 100 3 106/L. Within this group, CD4 counts were significantly higher in those with new cavitation (medians, 593 and 191 3 106/L, respectively; p 5 0.02) and pleural thickening (medians, 624 and 335 3 106/L, respectively; p 5 0.009) than in those without these features. There was no significant difference in the CD4 counts of smear-positive and smear-negative patients (medians, 457 and 335 3 106/L; p 5 0.67). Treatment outcomes by HIV group are shown in Table 5. There were no significant differences. No patient failed treatment and only two HIV-positive and one HIV-negative patients died (from Pneumocystis carnii pneumonia, Cryptococcus neoformans, and adenocarcinoma, respectively). The CD4 counts at start of treatment in the two patients with M. kansasii who died were lower than in those who survived the 12 TABLE 5 TREATMENT OUTCOMES IN HIV-POSITIVE AND HIV-NEGATIVE PATIENTS WITH NEW DEFINITE OR POSSIBLE M. kansasii DISEASE HIV-positive (n 5 18)

HIV-negative (n 5 35)

Treatment Outcome

n

%

n

%

Cured Died Transferred out

15 2 1

83 11 6

31 1 3

89 3 9

19

mo of treatment, but not significantly so (medians, 121 and 471 3 106/L, p 5 0.06).

DISCUSSION The findings from this prospective study of sputum NTM isolates from South African gold miners confirm that pulmonary NTM disease occurs with an unusually high incidence in this population. The incidence of new cases of NTM disease meeting ATS case-definitions was 37.0 and 8.8 per 100,000 employee-years for M. kansasii and M. scrofulaceum, respectively. NTM disease from other species was uncommon: there was one diagnosis of M. abscessus disease and no new cases of M. avium disease. M. kansasii has an association with mining and pneumoconiosis in other regions of the world (6–8), and pathogenicity in experimental animals can be increased by preexposure to coal and silica dust (9). Although M. scrofulaceum has rarely been described as a cause of pulmonary disease elsewhere (6), marked regional variation in the predominant pathogens is a general feature of NTM disease epidemiology (7). The unexpected predominance of M. scrofulaceum over other pulmonary pathogens such as M. avium could be a reflection either of local environmental exposure, regional variation in strain pathogenicity, or a particular association between mining and predisposition to pulmonary M. scrofulaceum disease. The effects of HIV injection on pulmonary NTM disease in miners are different from those described from other sites (7, 10–17). There was no apparent effect of HIV on NTM species distribution. Definite or possible pulmonary M. kansasii disease in HIV-positive miners mainly occurred at an early stage of HIV infection, as reflected by high CD4 counts and good short-term outcomes. The impact of HIV on radiologic features was less marked than reported in other populations (11– 17), although cavitation and pleural thickening were nonsignificantly less frequent in HIV-positive patients. In a larger retrospective case series of M. kansasii isolates from this site, cavitation was significantly less common in HIV-positive patients (18). An increasing effect on the radiologic appearances of HIV-associated M. tuberculosis disease with increasing degree of immunosuppression has been demonstrated (19, 20). The results from this study support a similar effect with M. kansasii disease in that CD4 counts were significantly lower in HIV-positive patients who had no cavitation or no pleural thickening than in HIV-positive patients with these features. A major question that arises from this study concerns the applicability of the ATS case-definitions for pulmonary NTM disease in populations with a high incidence of M. tuberculosis disease. The ATS case-definitions (3) were designed for use in the United States, but have been universally adopted (summarized in the APPENDIX). The possibility that patients may have been started on presumptive TB treatment is not directly considered, and the only potential for diagnosis of NTM disease on a single sputum isolate is in combination with lung granulomas on biopsy. Although obtaining NTM on culture of lung tissue provides strong evidence of NTM disease, and histology is useful in distinguishing colonization from indolent NTM disease, it is not possible to distinguish between different mycobacterial species on histologic grounds. In the current study, most patients were receiving presumptive TB treatment by the time a provisional NTM identification was available, and so would be expected to have a negative tissue culture had biopsy material been obtained. M. tuberculosis disease is common in this population, and would be the main differential in such cases. As such, demonstrating granulomas in lung tissue

20


would not help the diagnostic process if the biopsy culture was negative, and it could place undue weight on the significance of NTM isolates of low pathogenic potential. This problem is most clearly illustrated by patients with isolates of M. fortuitum, which has pathogenic potential but is the local common laboratory contaminant. Although several patients with M. fortuitum isolates did have strong clinical and radiologic evidence of mycobacterial disease, none failed presumptive TB treatment as would be expected in cases of genuine disease since M. fortuitum has intrinsic resistance to all first-line TB drugs. In addition, there is evidence that some of the M. kansasii and M. scrofulaceum isolates from patients who failed to meet the ATS case-definitions were pathogenically significant. Patients with M. kansasii and M. scrofulaceum isolates have a higher prevalence of NTM disease risk factors than do either trauma control subjects or patients with M. tuberculosis isolates at this site, suggesting that these isolates come from the patients themselves (2). It is not possible to comment on the clinical significance of NTM isolates from patients with M. tuberculosis coisolates, and these cases were not considered further. However, patients with pure M. kansasii isolates who did not meet the ATS case-definitions had a significantly higher prevalence of features suggestive of active mycobacterial disease than did patients with M. fortuitum isolates, and were significantly more likely to have been started on presumptive TB treatment before culture results were available. The treatment of M. kansasii disease is with first-line TB drugs and an early conversion to culture-negativity would be expected (8), making attempts to obtain further isolates less likely to be successful. Comparable data are not available for M. scrofulaceum isolates, but two patients with single isolates relapsed with M. scrofulaceum disease soon after stopping presumptive TB treatment. As has been reported from most studies attempting to apply case-definitions to patients with NTM isolates where initial investigations were conducted in routine clinical settings, a proportion of patients who failed to meet case-definitions were assessed on suboptimal information, either because less than three sputum cultures were taken initially or because of contamination of one or more cultures (10–16, 21–24). For example, a recent population-based study from the San Francisco area found that multiple specimens had been submitted for culture in only 66% of 270 patients from whom M. kansasii was isolated (24). Although this can be rectified by taking further specimens in untreated patients, confidently confirming or excluding NTM disease becomes impossible in those already receiving presumptive treatment. In this context, the ATS recommendation that liquid rather than solid culture media should be used to investigate possible NTM disease could only be consistently achieved before treatment is started if adopted for routine use as the initial mycobacterial culture medium, which would be prohibitively expensive for most countries, including South Africa. Because of the generally high clinical significance of M. kansasii isolates (11–16, 21–25), all patients with single M. kansasii isolates associated with clinical and radiologic disease were treated as if their isolate was pathogenic in the current study, whereas presumptive TB treatment was continued for comparable patients with other NTM species isolates. This practice of adopting different diagnostic thresholds for disease according to the NTM species deviates from that suggested by the ATS. However, no guidance is currently given for the interpretation of single sputum NTM isolates from patients who have warranted presumptive treatment. Starting presumptive TB treatment before isolate identification is common practice, especially in smear-positive patients: for example, in a trial treat-

VOL 160

1999

ment for M. kansasii disease based in the United Kingdom that included patients with two or more M. kansasii isolates, 149 of 173 patients (86%) were receiving TB treatment by the time of isolate identification (8). The number and management of patients excluded through having a single isolate only was not stated. This difficult area is one in which diagnostic certainty may be an impossible aim and in which the species of NTM isolated is of relevance. The choice in such cases lies between completing presumptive TB treatment, with a greater risk of recurrence if the isolate was the causative agent of disease, or changing to an NTM regimen, with an unnecessarily long duration, greater cost, and risk of side effects if the isolate was not clinically significant. Estimates of the clinical significance of sputum isolates vary greatly between NTM species (21, 23) and should be taken into consideration in such cases if overdiagnosis of disease caused by the less pathogenic species and underdiagnosis of disease caused by the more pathogenic species are to be avoided. Incorporating this subgroup of patients into NTM disease diagnostic guidelines would be of particular value for clinicians in countries such as South Africa where identification of multiple mycobacterial cultures is not routine practice because of the high incidence of TB and cost of speciation. There is a particularly strong case for reassessing diagnostic thresholds for pulmonary M. kansasii disease taking into account the growing number of case-series that document the potential for a rapidly fatal outcome with untreated HIV-associated disease (11–17). Each of these reports also include patients with single sputum isolates whose presumptive TB treatment was changed to that recommended for M. kansasii disease, implying that treating single sputum isolates has already become accepted practice in some specialist HIV units. The side-effect profiles and efficacy of NTM treatment regimens have improved markedly in recent years. This should perhaps be recognized by lowering diagnostic thresholds to recommend treatment of single sputum isolates of the more pathogenic NTM species in patients with clinical mycobacterial disease where further isolates cannot be obtained, especially if risk factors for NTM disease such as HIV-infection or chronic chest disease are also present. The acceptability of this approach would be improved if the shortest and simplest effective treatment regimens were established for the more common NTM pulmonary pathogens. Acknowledgment : The writers gratefully acknowledge Carol Van Blommestein, Themba Moyake, Victor Mchunu, Moses Dibe, and the Ernest Oppenheimer Hospital laboratory staff for their assistance with patient identification and follow-up.

References 1. Corbett, E. L., G. J. Churchyard, T. Moyake, P. Herselman, T. Clayton, B. Williams, D. Mulder, R. Hayes, and K. M. De Cock. 1998. HIV infection, silicosis and mycobacterial disease incidence in South African miners. Int. J. Tuberc. Lung Dis. 2:S301. 2. Corbett, E. L., G. I. Churchyard, T. Clayton, P. Herselman, B. Williams, R. Hayes, D. Mulder, and K. M. De Cock. 1999. Risk factors for pulmonary mycobacterial disease in South African gold miners: a casecontrol study. Am. J. Respir. Crit. Care Med. 159:94–99. 3. American Thoracic Society. 1997. Diagnosis and treatment of disease caused by nontuberculous mycobacteria. Am. J. Respir. Crit. Care Med. 156(Suppl.):S1–25. 4. Runyon, E. H. 1970. Identification of mycobacterial pathogens utilising colony characteristics. Am. J. Clin. Pathol. 54:578–586. 5. Kubica, G. P. 1973. Differential identification of mycobacteria: VII. Key features of identification of clinically significant mycobacteria. Am. Rev. Respir. Dis. 107:9–12. 6. Wolinsky, E. 1979. Nontuberculous mycobacteria and associated diseases. Am. Rev. Respir. Dis. 199:107–159.

Corbett, Blumberg, Churchyard, et al.: Nontuberculous Mycobacteria in Miners 7. Falkinham, J. O. 1996. Epidemiology of infection by nontuberculous mycobacteria. Clin. Microbiol. Rev. 9:177–212. 8. British Thoracic Society, Research Committee. 1994. Mycobacterium kansasii pulmonary infection: a prospective study of the results of nine months treatment with rifampicin and ethambutol. Thorax 49:442–445. 9. Policard, A., C. H. Gernez-Rieux, A. Tacquet, J. C. Martin, B. Devulder, and L. Le Bouffant. 1967. Influence of pulmonary dust load on the development of experimental infection by Mycobacterium kansasii. Nature 216:177–178. 10. Rigsby, M. O., and A. M. Curtis. 1994. Pulmonary disease from nontuberculous mycobacteria in patients with human immunodeficiency virus. Chest 106:913–919. 11. Lillo, M., S. Orengo, P. Ceroch, and R. L. Harris. 1990. Pulmonary and disseminated infection due to Mycobacterium kansasii: a decade of experience. Rev. Infect. Dis. 12:760–767. 12. Levine, B., and R. E. Chaisson. 1991. Mycobacterium kansasii: a cause of treatable pulmonary disease associated with advanced human immunodeficiency virus infection. Ann. Intern. Med. 114:861–868. 13. Carpenter, J. L., and J. M. Parks. 1991. Mycobacterium kansasii infections in patients positive for human immunodeficiency virus. Rev. Infect. Dis. 13:788–796. 14. Valainis, G. T., L. M. Cardona, and D. L. Greer. 1991. The spectrum of Mycobacterium kansasii disease associated with HIV-1 infected patients. J. Acquir. Immune Defic. Syndr. 4:516–520. 15. Bamberger, D. M., M. R. Driks, M. R. Gupta, M. C. O’Connor, P. M. Jost, R. E. Neihart, D. S. McKinsey, L. A. Moore, and the Kansas City AIDS Research Consortium. 1994. Mycobacterium kansasii among patients infected with human immunodeficiency virus in Kansas City. Clin. Infect. Dis. 18:395–400. 16. Witzig, R. S., B. A. Fazal, R. M. Mera, D. M. Mushatt, P. M. J. T. Dejace, D. L. Greer, and N. E. Hyslop. 1995. Clinical manifestations and implications of coinfection with Mycobacterium kansasii and human immunodeficiency virus. Clin. Infect. Dis. 21:77–85. 17. Campo, R. E., and C. E. Campo. 1997. Mycobacterium kansasii disease in patients infected with human immunodeficiency virus. Clin. Infect. Dis. 24:1233–1238. 18. Corbett, E. L., G. J. Churchyard, M. Hay, P. Herselman, C. Clayton, B. Williams, R. Hayes, D. Mulder, and K. M. De Cock. 1999. The impact of HIV infection on Mycobacterium kansasii disease in South African gold miners. Am. J. Respir. Crit. Care Med. 160:10–14. 19. Jones, B. E., S. M. M. Young, D. Antoniskis, P. T. Davidson, F. Kramer, and P. F. Barnes. 1993. Relationship of the manifestations of tuberculosis to CD4 cells counts in patients with human immunodeficiency virus infection. Am. Rev. Respir. Dis. 148:1292–1297. 20. Abouya, L., I. M. Coulibaly, D. Coulibaly, S. Kassim, A. Ackah, A. E. Greenberg, S. Z. Wiktor, and K. M. De Cock. 1995. Radiological manifestations of pulmonary tuberculosis in HIV-1 and HIV-2-infected patients in Abidjan, Cote d’Ivoire. Tuberc. Lung Dis. 76:436–440.

21

21. O’Brien, R. J., I. J. Geiter, and D. E. Snider. 1987. The epidemiology of nontuberculous mycobacterial disease in the United States: results from a national survey. Am. Rev. Respir. Dis. 135:1007–1014. 22. Evans, S. A., A. Colville, A. J. Evans, A. J. Crisp, and D. A. Johnston. 1996. Pulmonary Mycobacterium kansasii infection: comparison of the clinical features, treatment and outcomes with pulmonary tuberculosis. Thorax 51:1248–1252. 23. Choudhri, S., J. Manfreda, J. Wolfe, S. Parker, and R. Long. 1995. Clinical significance of nontuberculous mycobacterial isolates in a Canadian tertiary care center. Clin. Infect. Dis. 21:128–133. 24. Bloch, K. C., L. Zwerling, M. J. Pletcher, J. A. Hahn, J. L. Gerherding, S. M. Ostroff, T. J. Vulga, and A. I. Reingold. 1998. Incidence and clinical implications of isolation of Mycobacterium kansasii: results of a 5 year, population based study. Ann. Intern. Med. 129:698–704. 25. Ahn, C. H., J. W. McLarty, S. S. Ahn, I. Ahn, and G. A. Hurst. 1982. Diagnostic criteria for pulmonary disease caused by Mycobacterium kansasii and Mycobacterium intracellulare. Am. Rev. Respir. Dis. 125:388– 391.

APPENDIX American Thoracic Society Case-Definitions for Pulmonary NTM Disease

• Compatible signs/symptoms and reasonable exclusion of other disease to explain condition (including M. tuberculosis disease) • Pulmonary changes compatible with NTM disease on chest radiography and/or high resolution CT scanning (infiltrate/multiple nodules/cavities/multifocal bronchiectasis) • plus any one of: —two NTM isolates from separate sputum or bronchoalveolar lavage (BAL) specimens and one smear-positive specimen or —three NTM isolates from separate sputum or BAL specimens if smear-negative or —heavy growth and numerous AFB on a single available BAL specimen or —single NTM isolate from lung tissue (transbronchial or open lung biopsy) with no other cause found or —lung biopsy tissue showing granulomatous inflammation or AFB plus a single available NTM isolate from sputum or BAL