Nontuberculous Mycobacterial Disease in Children - Epidemiology, Diagnosis & Management at a Tertiary Center

Abstract

Background: There are limited data on the epidemiology, diagnosis and optimal management of nontuberculous mycobacterial (NTM) disease in children.

Methods: Retrospective cohort study of NTM cases over a 10-year-period at a tertiary referral hospital in Australia.

Results: A total of 140 children with NTM disease, including 107 with lymphadenitis and 25 with skin and soft tissue infections (SSTIs), were identified. The estimated incidence of NTM disease was 0.6-1.6 cases / 100,000 children / year; no increasing trend was observed over the study period. Temporal analyses revealed a seasonal incidence cycle around 12 months, with peaks in late winter/spring and troughs in autumn. Mycobacterium-avium-complex accounted for most cases (77.8%), followed by Mycobacterium ulcerans (14.4%) and Mycobacterium marinum (3.3%). Polymerase chain reaction testing had higher sensitivity than culture and microscopy for acid-fast bacilli (92.0%, 67.2% and 35.7%, respectively). The majority of lymphadenitis cases underwent surgical excision (97.2%); multiple recurrences in this group were less common in cases treated with clarithromycin and rifampicin compared with clarithromycin alone or no anti-mycobacterial drugs (0% versus 7.1%; OR:0.73). SSTI recurrences were also less common in cases treated with two anti-mycobacterial drugs compared with one or none (10.5% versus 33.3%; OR:0.23).

Conclusions: There was seasonal variation in the incidence of NTM disease, analogous to recently published observations in tuberculosis, which have been linked to seasonal variation in vitamin D. Our finding that anti-mycobacterial combination therapy was associated with a reduced risk of recurrences in patients with NTM lymphadenitis or SSTI requires further confirmation in prospective trials.

Fig 1. Seasonal distribution of nontuberculous mycobacterial lymphadenitis and skin and soft tissue infections over the study period according to the month of initial presentation.

Fig 2. Time series of nontuberculous mycobacterial lymphadenitis and skin and soft tissue infections combined over the study period.

The lower plot shows the raw data; the upper plot shows the 3^rd level denoised data. Denoising was performed using a discrete wavelet transform with Daubechies length 6 wavelet filter.

Fig 3. Power spectrum obtained from a continuous wavelet transform based on a Morlet wavelet filter for the detrended and denoised time series of nontuberculous mycobacterial lymphadenitis and skin and soft tissue infection cases combined.

The graph shows the power for each period or cycle (y-axis) along the time series (x-axis). High values are shown in dark-red and low values in dark-blue. The dark line contains the statistically significant periods (p<0.05); the white line is the cone of influence that delimits the area not influenced by edge-effects.