Nontuberculous Mycobacteria, Botswana, 2011–2014

We documented a 6-fold increase in the frequency of nontuberculous mycobacteria isolated from clinical samples in Botswana during 2011–2014. Because antituberculosis treatment is often initiated only on the basis of acid-fast bacilli smear-positive microscopy results, some patients with nontuberculous mycobacterial infections might have received inappropriate treatment.

The proportions of M. tuberculosis and NTM cases during the study period were comparable (5.6% vs. 5.5%). This finding creates a high possibility of misdiagnosis and inappropriate treatment of M. tuberculosis/NTM cases. Most (74.5%) NTM samples were sputum; 22% were gastric aspirates, and the remaining 3.4% were obtained from other body sites. Of specimens from which NTM was isolated, 53.3% were from male patients and 43.4% were from female patients; for 3.1%, patient sex was not captured. The first specimen submitted per patient was analyzed. From the analysis we removed duplicate samples that might have resulted from submission of samples for follow-up from the same patient.
One third (33.5%) of samples from which NTM was isolated were from patients 0-14 years of age, followed by 27.3% from patients 35-54 years of age ( Figure 1). For at least 207 (10.4%) of the 1,999 NTM samples, the age of the patient was not captured.
Overall, the proportion of isolated NTM . Seventy-eight (24.0%) isolates could be identified only as Mycobacterium spp. because of a limitation of the assay used. Seventy-four (46.0%) patients in whom M. intracellulare was isolated were also HIV positive, suggesting that NTM might be an important opportunistic pathogen in HIV infection. M. avium-intracellulare has also been reported to cause most pulmonary infections in several studies (8,10,11). These reports are useful because failure to speciate NTM isolates or conduct drug susceptibility testing might be overlooking an emerging pathogen with potential to cause illness and even death in HIV-infected persons.
Most samples were from male patients, a finding consistent with findings by Moore et al. (1). This finding has DISPATCHES been attributed to risk factors such as increased rates of smoking and chronic obstructive pulmonary disease among men (1). However, investigators in the United States (3) observed a female predominance.

Conclusions
We observed a high occurrence of NTM in patients 35-54 years of age. We also noted an unexpected increase of >30% in reported NTM cases in children <14 years of age, which might represent a peculiar childhood vulnerability to NTM infections.
According to the Botswana National Tuberculosis Programme Manual, no information is available in Botswana about the prevalence of disseminated NTM (12). Our study provides basic but crucial data that may stimulate discussion toward policy change in the laboratory investigation and treatment of NTM isolates for better patient care.
The change in 2011 in the use of media from Lowenstein-Jensen to MGIT (Beckton-Dickinson, https://www. bd.com) culture might have contributed to an increase in NTM identification. A similar study by Chihota et al. (13) yielded more NTM after such a change. However, if the increase in this study resulted from a change in media, only the first year would have recorded an increase, followed by a stabilizing trend in subsequent years. Therefore, the increasing trend we observed reflects a convincing climb in NTM cases. Furthermore, our results suggest that patients with NTM might have received unnecessary TB treatment (7). This finding underscores the need to unequivocally identify mycobacterial isolates because NTM appears to be more frequently encountered (9).
Our study has limitations. Our results might not reflect the prevalence of NTM in the country because the samples analyzed comprised only persons suspected to have TB. This fact notwithstanding, the National Tuberculosis Referral Laboratory is the sole laboratory in Botswana that performs culture and thus represents a reasonable sampling of NTM countrywide among persons suspected to have TB infection. The prevalence of NTM might therefore be higher than that reported in this study.
We propose that mycobacterial isolates be routinely speciated and subjected to drug susceptibility testing. We base this proposal on our observation that NTM is isolated at nearly the same frequency as M. tuberculosis and is isolated predominantly in children <14 years of age. Treatment of NTM as M. tuberculosis underestimates the clinical significance of NTM and is likely to negatively affect treatment outcomes and inflate reports of M. tuberculosis.

Acknowledgments
We thank the Botswana Ministry of Health and Wellness for allowing access to the data used in this study.

About the Author
Ms. Mbeha is a Laboratory Scientific Officer at the National Tuberculosis Reference Laboratory in Gaborone, Botswana. Her primary research interests include tuberculosis pathogenesis and treatment effectiveness.