Keywords
Mycobacterium abscessus treatment -
Mycobacterium abscessus
-
Mycobacterium massiliense
-
Mycobacterium bolletii
- bronchiectasis - cystic fibrosis
Mycobacterium abscessus complex (MABc) is the second most common nontuberculous mycobacteria (NTM) grouping
causing human disease after Mycobacterium avium complex (MAC).[1 ]
[2 ] MABc causes a wide spectrum of human disease, most commonly pulmonary disease, although
it can cause soft tissue disease, bone disease, and disseminated disease in immunocompromised
hosts.[1 ]
[3 ] The spectrum of infection severity is heterogeneous, ranging from asymptomatic colonization
to progressive and mortal disease.[4 ]
[5 ] Unfortunately, MABc is a highly drug-resistant organism and few oral antibiotics
show in vitro activity, making long-term treatment of this infection extremely complicated.[1 ]
[4 ]
[5 ] In this article, we will review the treatment of MABc pulmonary disease as well
as its epidemiology, disease manifestations, and aspects of its diagnosis that are
germane to disease management.
Microbiology and Epidemiology
Microbiology and Epidemiology
Mycobacterium abscessus Complex Taxonomy
MABc is one of a group of NTM “rapid growers,” which includes Mycobacterium fortuitum complex, Mycobacterium chelonae , and MABc, all having the characteristic of growing within 7 days on solid media
subculture (faster in liquid media) as well as the ability to grow in standard blood
culture.[1 ]
[5 ] It was only in 1992 that MABc organisms could be identified as distinct from M. chelonae and thus these organisms are undifferentiated in literature published before that
time.[1 ]
[5 ]
[6 ] As recently as 2006, M. abscessus was reclassified to represent a complex containing three subspecies: M. abscessus , Mycobacterium massiliense , and Mycobacterium bolletii .[1 ]
[5 ] Over the subsequent 7 years, the taxonomic and nomenclature classifications for
these organisms have undergone additional changes resulting in the current understanding
of MABc ([Fig. 1 ]).[5 ] It can be difficult to analyze literature regarding “M. abscessus ” treatment as many studies do not separately analyze these subspecies, which have
differential drug susceptibilities and accordingly differential treatment success.
Additionally, since early studies did not differentiate between M. abscessus and M. chelonae , it is difficult to judge the results of these studies, as M. chelonae is often a colonizer or contaminant.[1 ]
Fig. 1 Changes over time to subspecies organization of MABc. (Adapted from Lee et al.[5 ])
Organism Microbiology
Several microbiologic features of MABc are important to highlight because of their
implications for disease management. First, like many NTM species, these organisms
are ubiquitous in soil and water environments.[1 ]
[7 ]
[8 ] As a result, patients with an underlying predilection to initial infection (e.g.,
bronchiectasis, other structural lung disease) have a high risk of reinfection after
adequate treatment, as well as a risk of infection with different strains of the same
organism.[9 ]
[10 ] Second, the organisms have a propensity for biofilm formation. This protects them
not only from the immune response but also from antibiotics, both of which make true
eradication very difficult.[8 ]
[11 ] Third, the presence or absence of a functional erm gene (RNA methylase gene), discussed in more detail in next sections, has important implications
for drug resistance to macrolide antibiotics. Since this drug class constitutes the
most effective oral antibiotic for susceptible NTM organisms, this biology is crucial
to treatment choice and response.[12 ]
[13 ]
[14 ] Fourth and least well understood, there appears to be other genotypic variations
within the MABc subspecies that influence disease manifestations and treatment response.[15 ]
[16 ]
[17 ]
Disease Epidemiology
Along with most other NTM infections, the incidence and prevalence of MABc pulmonary
infections appears to be increasing, although it is unclear what role increased awareness
and improved diagnostics is playing in this process.[18 ]
[19 ]
[20 ] Of the complex subspecies (subsp), subsp abscessus is generally the most common, although there are notable geographic variations in
this finding with subsp massiliense even more prevalent in some regions ([Table 1 ]).[4 ] Since not all laboratories can perform subspecies identification and the process
of getting results from a reference laboratory can take many weeks, knowledge of local
epidemiology can be helpful in crafting empiric antibiotic regimens if they are needed
while waiting further subspecies identification. That said, since subsp abscessus is the most drug-resistant of the group and since in no region is the prevalence
low enough to ignore, empiric treatment of MABc should consider subsp abscessus .
Table 1
MABc propotions of M. abscessus , M. massiliense , M. bolletii
Study
Location
Total number
M. abscessus (%)
M. massiliense (%)
M. bolletii (%)
Zelazny, 2009
[143 ]
United States
40
67.5
27.5
5
van Ingen, 2009
[22 ]
Netherlands
39
64
21
15
Roux, 2009
[24 ]
France
50
60
22
18
Harada 2012
[144 ]
Japan
102
71
26
3
Yoshida 2013
[36 ]
Japan
143
63
35
2
Nakanaga, 2014
[35 ]
Japan
115
60
37
3
Huang, 2013
[105 ]
Taiwan
79
43
56
1
Kim, 2008
[145 ]
South Korea
126
53
45
2
Koh, 2011
[23 ]
South Korea
158
44
55
1
Lee, 2014
[146 ]
South Korea
404
50
49
1
Note: Adapted from Koh et al.[4 ]
Another nuance in the evaluation of MABc literature and clinical management of the
disease is the fact that in the respiratory tract, culture positivity (infection)
does not inherently mean disease.[1 ]
[21 ] In the lungs, all subspecies can behave in a nonpathogenic way, essentially “colonizing”
the airway with minimal clinical impact.[1 ]
[21 ]
[22 ] Other times, the organisms can cause airway inflammation without causing parenchymal lung disease. In both situations, infection can become progressive
within the airway and surrounding lung parenchyma, ultimately causing disease. Sputum
culture positivity does not definitively indicate which situation is true in a patient
at any given moment.[1 ] This has implications for treatment but also for interpreting literature in which
it is not always clear whether patients included in studies have “colonization” versus
“disease,” which can change reporting on manifestations and treatment response; therefore,
this word of caution must be kept in mind when interpreting published literature.
Clinical Features of Mycobacterium abscessus Complex Pulmonary Disease
Clinical Features of Mycobacterium abscessus Complex Pulmonary Disease
An overwhelming majority of the cases of MABc pulmonary infection and disease occur
in hosts with underlying lung disease.[9 ]
[15 ]
[22 ]
[23 ] This can be airway disease such as bronchiectasis, which allows accumulation of
and prevents clearance of environmental colonizing microorganisms, or parenchymal
disease such as emphysema or pulmonary fibrosis, which by causing destruction and
impaired circulation at the tissue level inhibits immune surveillance against and
clearance of foreign microbes. This susceptibility to infection is particularly pronounced
in patients with cystic fibrosis (CF), who have bronchiectasis but also airway ciliary
deficiencies creating a unique clinical phenotype covered in detail later in this
review.[24 ]
[25 ]
[26 ] Because all the above physiology allows other organisms to gain easier access to
the respiratory tree as well, individuals with MABc pulmonary infection are usually
colonized with and often infected with other pathogens.[9 ]
[27 ] Treatment sometimes needs to take these organisms into account or clinicians need
to consider a superinfection with these other organisms as one of the possible reasons
for worsening in a MABc patient on therapy. There is a poorly understood interplay
of the lung microbiome here, and in some patients, treatment of MABc allows other
pathogens to flare (and vice versa) as organisms compete for resources within the
airway.[28 ]
[29 ]
MABc pulmonary infection typically presents in an indolent fashion with waxing and
waning symptoms, and many patients do not have significant progression for months
or years after infection.[1 ]
[21 ]
[30 ]
[31 ] In light of this, treatment initiation is not always indicated even if disease criteria
described below is met, and when it is indicated, this is rarely urgent.[1 ] Like MAC, the presentation is typically scattered nodular infiltrates in or around
areas of structural lung disease.[1 ]
[9 ]
[15 ] Also like MAC, less commonly the disease can be cavity-forming or superinfect structural
lung cavities, and it is not clear whether or not cavity formation is more common
with MABc or MAC.[15 ]
[32 ] Because of the higher mycobacterial burden and less optimal blood supply within
these areas of cavitary physiology, the presence of cavities has implication for treatment
decisions.[1 ] Coinfection with MAC is not uncommon and MABc can be seen in up to 50% of pulmonary
MAC cases in some series.[9 ]
[30 ] An estimated 15 to 30% of MABc cases have MAC isolated from their sputum as well.[33 ] Coinfection presents a clinical conundrum as it can be difficult to ascertain which
organism is playing a pathogenic role (or if both are) and should be targeted for
therapy.[1 ]
[33 ]
Diagnosis of Infection versus Disease
Diagnosis of Infection versus Disease
The identification of MABc is still largely culture-based, although new technologies
are revolutionizing laboratory-based diagnostics in this area.[34 ]
[35 ] However, microbe identification (infection) does not equal disease, especially in
pulmonary specimens, so pulmonary disease diagnosis is worth reviewing in any discussion
of treatment. American Thoracic Society guidelines for pulmonary NTM disease are listed
below:[1 ]
Symptoms attributable to an infection with the organism.
At least two positive sputum cultures or one positive bronchoalveolar lavage or lung
biopsy culture with the organism.
Radiographic changes compatible with infection with the organism.
In addition to the potential of nonpathogenic colonization, organisms in MABc can
occasionally be contaminants of the laboratory process because of their prevalence
in nonsterile environments; thus, as with most infectious diseases, isolation from
a site without clinical evidence of pathology should be viewed with suspicion.[1 ]
[7 ]
Drug Resistance and Susceptibility Testing
Drug Resistance and Susceptibility Testing
As mentioned previously, organisms in the MABc are highly drug resistant and understanding
their resistance properties and laboratory ways to identify them is a crucial part
of treatment.
Slow Growth
Compared with most other pathogens, even rapid-growing mycobacteria are slow growing.
This property creates treatment difficulties for multiple reasons. Since many antibiotics
work at least partly on dividing pathogens, slow cell replication means slower rates
of antibiotic activity. Of the antibiotics typically used to treat MABc, this property
limits the efficacy of β-lactams (BLs) and means most of the antibiotic classes used
rely on targeting ribosomal protein synthesis, which is not dependent on bacterial
division (macrolides, tetracyclines, aminoglycosides, fluoroquinolones, oxazolidinones).[36 ] The slow growth of the organisms also makes culture-based susceptibility testing
time-consuming and difficult.
Waxy Cell Wall Barrier
The cell wall of MABc has a high lipid content, creating a waxy barrier that is difficult
for antibiotics to penetrate.[37 ] This fact alone is potentially sufficient to confer resistance to many BLs and likely
impairs susceptibility to aminoglycosides.[38 ] Even when not completely impermeable, by lowering the antibiotic concentration within
the bacteria, the cell wall plays a role in allowing other resistance mechanisms to
be protective when they would otherwise be overwhelmed by high enough antibiotic concentrations.[39 ]
Biofilm Formation
Biofilm formation is protective for many microbial organisms by creating a surrounding
environment that is poorly penetrated by the immune system and antibiotics. Biofilms
also create an environment within which bacteria can survive despite less metabolic
activity, rendering them more resistant not only to antibiotics that target cell division
but also to protein synthesis and adenosine triphosphate (ATP) synthesis processes.[11 ]
[40 ] MABc has been described to form these biofilms in airways, alveoli, and pulmonary
cavities, which likely explains part of its resistance to treatment and the difficulty
with infection eradication.[11 ]
[41 ]
[42 ]
Drug-Neutralizing Enzymes and Export Systems
There are many neutralizing enzymes and drug export systems that enhance the ability
of MABc to survive antibiotic exposure ([Table 2 ]), including an adenosine diphosphate (ADP)-ribosyltransferase and mono-oxygenase
that may confer the complexes known rifampin resistance.[36 ]
[43 ] The most importance of these resistance mechanisms is an inducible erm gene which is present in some members of the MABc and confers macrolide resistance
that is expressed upon macrolide exposure.[44 ] This mechanism is notable both for the important role macrolides play in management
of NTM infections and because of the difficulty in identifying its presence. Since
it is an inducible enzyme, the mechanism will not be picked up on in vitro antibiotic
susceptibility testing unless this testing is done after incubating the organism in
the presence of macrolides for 14 days.[1 ]
[44 ] As a result, the Clinical and Laboratory Standards Institute (CLSI) recommends that
macrolide testing for this complex include prolonged incubation testing for inducible
macrolide resistance.[1 ]
[45 ] Additionally, the erm gene can be present and inactive, so molecular tests that only probe for the presence
of the gene do not conclusively prove macrolide resistance.[14 ]
[36 ] Since many laboratories are not capable of performing prolonged drug incubation
and do not have the molecular diagnostics to either identify the erm gene or identify if it is active, accurate macrolide susceptibility testing may require
sending the isolate to a mycobacterial-specific reference laboratory. Labs that report
macrolide susceptibility without performing appropriate testing may provide results
that mislead the untrained provider leading to suboptimal treatment regimens. An active
erm gene is present in most MABc subsp abscessus isolates, in some subsp bolletii , and in a small proportion of subsp massiliense .[12 ]
[13 ]
[46 ]
Table 2
Possible mechanisms of resistance of MABc
Antibiotic
Locus and genes
Proteins involved
Resistance mechanism
Hydrophilic antibiotics
–
–
Selective permeability of cell envelope
Aminoglycosides
MAB_4395
Aminoglycoside 2-N-acetyltransferase
Antibiotic-modifying enzymes
MAB_0327
Aminoglycoside phosphotransferases
MAB_0951
MAB_3637c
MAB_4910c
MAB_4395
Rifampicin
MAB_0951
Rifampicin ADP-ribosyltransferase
−
Beta-lactams
MAB_2875
Beta-lactamase
Antibiotic-degrading enzymes
Macolides
erm(41) gene
23S RNA methyltransferase
Target-modifying enzymes
MAB_2297
Several antibiotics
scattered in genome
ABC transporters. Mmpl. family
Efflux pumps
Note: Adapted from Nessar et al.[36 ]
Genetic Polymorphisms
Because of specific genetic polymorphisms within the MABc, these organisms are less
susceptible to antibiotics than most other mycobacteria (higher minimum inhibitory
concentrations, MICs). Classic in this regard are a couple base-pair substitutions
in the embB ethambutol resistance determining region which confer high-level (MICs > 64 mg/L)
ethambutol resistance to all members of the MABc.[47 ] Although the MABc is not intrinsically resistant to fluoroquinolones, only a couple
point mutations are needed in DNA gyrase subunits A and B to confer fluoroquinolone
resistance (MICs > 8 mg/L).[48 ] This combined with the frequent use of fluoroquinolones for a variety of medical
conditions, and almost ubiquitous use in agriculture, means that a sizeable proportion
of MABc organisms are quinolone-resistant even if the patient has not received prior
quinolone therapy for their mycobacterial disease.[49 ]
[50 ]
[51 ] Aminoglycoside resistance can also develop more easily than with other bacteria
as MABc only possesses a single copy of the ribosomal RNA (rRNA) operon that aminoglycosides
target.[36 ]
[52 ] This allows several single genetic mutations to cause high-level (MICs > 1,000 mg/L)
drug resistance to aminoglycosides.[36 ]
[52 ] In addition to the inducible resistance mechanism to macrolides mentioned above,
there is also the potential for a point mutation in the 23S rRNA peptidyl transferase
region of MABc, which confers macrolide resistance (MICs > 4 mg/L), most commonly
in the setting of prior macrolide monotherapy.[53 ]
Susceptibility Testing
The CLSI currently recommends the broth microdilution minimal inhibitory concentration
(MIC) method for determining the susceptibility of MABc using a panel of 10 antimicrobials
([Table 3 ]).[45 ] Agar tests, including the E-test, are not recommended because of inconsistency of
results.[54 ] It is important to note that there is no proven clinical correlation between MICs
and treatment outcome for pulmonary MABc.[1 ] There does appear to be a correlation between in vitro susceptibility and clinical
response in skin and soft tissue infection, although this observation has not been
prospectively confirmed.[1 ] In general, amikacin, cefoxitin, imipenem, clofazimine, and macrolides (the last
only in MABc organisms without an active erm gene) have the most reliably low MICs suggesting the highest chance of clinical activity,
although with no antibiotic is this reliable enough to dispense with MIC testing ([Table 4 ]).[1 ]
[36 ]
Table 3
CLSI MIC Breakpoints for MABc
Antimicrobial
MIC for broth dilution (μg/mL)
Susceptible
Intermediate
Resistant
Amikacin
≤16
32
≥64
Cefoxitin
≤16
32–64
≥128
Clarithromycin
≤2
4
≥8
Ciprofloxacin
≤1
2
≥4
Doxycycline
≤1
2−8
≥16
Imipenem
≤4
8
≥16
Linezolid
≤8
16
≥32
Moxifloxacin
≤1
2
≥4
Trimethoprim-sulphamethoxazole
≤2/38
n/a
≥4/76
Tobramycin
≤4
8
≥16
Note: Adapted from National Committee for Clinical Laboratory Standards (CLSI). Susceptibility
testing of mycobacteria, nocardiae, and other aerobic actinomycetes. Approved Standard.
Wayne, PA: NCCLS; 2011. Document No. M24-A2.[45 ]
Table 4
Antibiotic susceptibility of MABc as defined by MIC (mg/L)
Antibiotic
Studies
Number of subjects
MIC range (mod) (mg/L)
Percent susceptible[a ]
Clarithromycin
2
48–74
0.03–16 (0.03)
83–99
Cefoxitin
2
48–74
16–128 (32)
11–99
Imipenem
2
48–74
1–64 (8)
8–55
Ciprofloxacin
2
48–74
0.016–8 (2)
44–57
Levofloxacin
1
21
8–64 (32)
Not reported
Moxifloxacin
1
21
2–32 (16)
73
Doxycycline
1
20
2 - >128 (>128)
5
Minocycline
1
20
0.25 - >64 (>64)
5
Tigecycline
1
20
≤0.06 - 1 (0.12)
100
Linezolid
1
98
0.5 - 128 (32)
23
Sulfamethoxazole
2
48–74
4 - 256 (256)
1–12
Amikacin
2
48–117
0.25 - >128 (2 and 16)
87–94
Tobramycin
3
21–117
4 - >128 (8 and 16)
36–95
Clofazimine
1
117
0.25-1 (0.5)
99
Note: Adapted from Nessar et al.[36 ]
a
Susceptibility breakpoints as defined by Griffith et al
1 and Woods et al.
[54 ]
Treatment Choice and Treatment Response
Treatment Choice and Treatment Response
Because of a paucity of prospective, controlled, or randomized treatment studies,
most treatment recommendations for MABc disease rely heavily on small retrospective
or observational datasets, extrapolation from the treatment of other mycobacterial
organisms, in vitro studies, and expert opinion.[1 ] In this context, individual patient response, treatment toxicity, and overall goals
of therapy are important considerations. That said, owing to the disease complexity
and use of uncommon antimicrobials, expert recommendations and guidelines have an
important role to play in management, a sobering concept given how infrequently even
basic guidelines are followed for NTM disease.[1 ]
[55 ] Below we have attempted to outline treatment considerations and recommendations
for pulmonary MABc as well as specific considerations for the various antibiotics
most frequently used.[56 ] There are some general principles that should be followed, including prolonged therapy,
an induction phase of therapy with a three- to four-drug regimen including one to
two different active intravenous (IV) agents, and a suppressive phase which should
involve at least two oral or inhaled antibiotics considered active based on drug susceptibilities,
since most patients cannot tolerate 12 to 18 months of IV therapy.[1 ]
Pulmonary MABc disease is difficult to manage and treatment outcomes are generally
poor.[1 ]
[4 ] There is a clear difference in treatment response based on subspecies with subsp
massiliense having significantly better treatment outcomes than subsp abscessus ([Table 5 ]). This is almost certainly due to the ability to use macrolide-based therapy for
massiliense , which allows a more effective overall regimen and less treatment-related toxicity.[1 ]
[57 ]
[58 ] IV amikacin is generally the most active agent and is recommended as part of most
treatment regimens if feasible from a renal and ototoxicity standpoint using three
times per week (TIW) dosing.[1 ]
[4 ] Generally, an IV BL (either imipenem or cefoxitin) should be added for the initial
phase of therapy as well.[1 ] Imipenem has a better toxicity profile than cefoxitin and older concerns about its
efficacy have been somewhat refuted by newer in vitro data, so it is our preferred
BL in most cases.[36 ]
[59 ] At least 4 to 8 weeks of dual IV induction therapy is recommended to give the best
chance of prolonged clinical response.[1 ]
[15 ]
[58 ] In macrolide-susceptible isolates, an oral macrolide should be part of the treatment
regimen, but even in nonsusceptible isolates, there may be a role for an oral macrolide
given the benefits to airway inflammation and against other concurrent potential pathogens
(see section Macrolides below). Clofazimine is often the only other oral antibiotic
with a favorable susceptibility profile, although there are no rigorous clinical trials
proving its efficacy.[60 ]
[61 ] However, in vitro data and small case series suggest it could be a useful agent
to add instead of or along with a macrolide.[60 ]
[61 ] Given how difficult it is to obtain the recommended 12 months of negative sputum
cultures for MABc lung disease, symptom and radiographic improvement are useful markers
of treatment success.[1 ]
Table 5
Studies of antibiotic treatment response in MABc subspecies abscessus and massiliense
Study, year, country
Study number and population
Study design
Study period
Treatment duration (wk)
Treatment regimen
Treatment success (%)[a ]
Subsp abscessus
Choi et al, 2017, South Korea[140 ]
12 patients with acquired macrolide resistance starting treatment
Retrospective review
2006–2016
Average: 96 (64–176)
IV amikacin + IV cefoxitin or imipenem + oral macrolide ± FQ/TCN/CFZ
0
(1 converted s/p surgery)
Yang et al, 2017, South Korea[61 ]
42 patients starting clofazimine for initial (36%) or salvage (64%) treatment
Retrospective review
2013–2015
Median: 48 (25–48)
Oral CFZ + oral macrolide (85% resistant) + IV amikacin (67%) and/or IV BL (67%)
24
(81 had improved sx)
Park et al, 2017, South Korea[57 ]
19 patients starting treatment after disease progression
Retrospective review
2006–2015
Median: 61 (29–116)
• Amikacin: 16 (2–50)
• Cefoxitin: 2 (0–6)
• Imipenem:3 (0–5)
At least two IV agents (amikacin, imipenem, or cefoxitin) + oral macrolide
26
(surgery in 3 patients)
Koh et al, 2017, South Korea[58 ]
67 patients initiating therapy
Prospective observational cohort
2002–2012
All: 48
• All: Initial IV therapy 4 wk
IV agent (amikacin or cefoxitin) + macrolide ± quinolone ± doxycycline
42
(surgery in 6 responders and in 3 nonresponders)
Lyu et al, 2014, South Korea[141 ]
26 patients initiating therapy
Retrospective review
2006–2012
Average: 65 (SD ± 46)
• IV therapy mean: 39 (SD ± 16)
IV amikacin ± IV imipenem or cefoxitin + oral macrolide
42
(surgery in 8 patients)
Kim et al, 2012, South Korea[15 ]
24 patients initiating therapy
Retrospective review
2004–2009
All: 96
• All: initial IV therapy 4 wk
IV amikacin and IV cefoxitin + oral clarithromycin + oral ciprofloxacin ± oral doxycycline
50
(33 had CT-based imaging improvement)
Subsp massiliense
Park et al, 2017, South Korea[57 ]
17 patients starting treatment after disease progression (out of larger group of 56
patients)
Retrospective review
2006–2015
Median 97 (64–120)
• Amikacin: 7 (2–30)
• Cefoxitin: 3 (0–6)
• Imipenem: 0 (0–4)
At least two IV agents (amikacin, imipenem, or cefoxitin) + oral macrolide
82
(surgery in 2 patients)
Koh et al, 2016, South Korea[142 ]
71 patients initiating therapy
Prospective observational cohort
2007–2012
Median 96 (92–97) for 2007–2010
• All: IV agent 4 wk
and
Median 61 (51–72) for 2010–2012
• All: IV agent 2 wk
IV amikacin and IV imipenem or cefoxitin + macrolide + quinolone (if 2007–2010 )
98
(96 had symptom improvement)
Lyu et al, 2014, South Korea[141 ]
26 patients initiating therapy
Retrospective review
2006–2012
Average: 48 (SD ± 18)
IV therapy mean: 19 (SD ± 10)
IV amikacin ± IV imipenem or cefoxitin + oral macrolide
96
(82 had symptom improvement, surgery in 1 patient)
Kim et al, 2012, South Korea[15 ]
34 patients initiating therapy
Retrospective review
2004–2009
All: 96
• All: initial IV therapy 4 wk
IV amikacin and IV cefoxitin + oral clarithromycin + oral ciprofloxacin ± oral doxycycline
100
(88 had CT-based imaging improvement)
Abbreviations: CFZ, clofazimine; CT, computed tomography; FQ, fluoroquinolones; IV,
intravenous; s/p, status post; sx, symptoms; TCN, tetracycline.
a Treatment success defined as sustained sputum culture conversion.
Because of the poor response to antibiotic treatment, in many cases surgical intervention
in combination with antibiotic therapy holds the greatest chance of prolonged remission
or cure, especially in cases where the areas of pulmonary infection are focal.[62 ]
[63 ] Although most NTM lung surgery data are in pulmonary MAC, the surgical technique
is similar for MABc.[64 ]
[65 ] In pulmonary MAC, surgical intervention has been shown to be safe and effective,
although there is not a large enough body of literature in MABc to prove it is equally
safe.[64 ]
[65 ] If surgery is pursued, we recommend aggressive antimicrobial therapy in the 4 to
8 weeks before surgery to lower the bacterial burden and geographic extent of infection
before lung resection. We also strongly recommend this take place at a center surgically
experienced in NTM lung disease.
Specific Antimicrobials
One of the key components of MABc treatment is the use of three or more antimicrobials
in most treatment regimens to increase efficacy and decrease the development of antibiotic
resistance. Thus, the interplay between different antibiotics is an important dynamic
with synergy or enhanced abilities to protect each other from the development of resistance
as key theoretical components of more effective regimens.
Beta-lactams
Although BLs are a mainstay in the treatment of many gram-positive bacterial infections,
most of the class has little utility against MABc because of its production of BlaMABc , a broad-spectrum β-lactamase that inactivates most BLs.[66 ] Unfortunately, BlaMABc is not effectively inhibited by the standard β-lactamase inhibitors clavulanate,
tazobactam, and sulbactam.[66 ] Imipenem and cefoxitin are more slowly hydrolyzed than other BLs, allowing them
to maintain activity against many MABc isolates.[66 ] Other carbapenems such as meropenem are not as active as imipenem, and in the treatment
of MABc, imipenem is the carbapenem of choice.[67 ] In vitro studies suggest that imipenem has lower relative MICs than cefoxitin, although
there can be difficulties around the lack of reliability of imipenem MICs.[1 ]
[59 ]
[68 ] Interestingly, in animal models, the new β-lactamase inhibitor, avibactam, effectively
inhibits BlaMABc potentially improving imipenem efficacy against these organisms and significantly
reducing MICs against the new BL ceftaroline, although not enough to make ceftaroline
useful in treatment.[69 ]
[70 ]
[71 ] There is potential intracellular synergy between imipenem and amikacin that is not
present with cefoxitin and amikacin, another argument for imipenem over cefoxitin
since both drugs are often used in combination with an aminoglycoside.[1 ]
[59 ]
[68 ] Although there are no comparative studies to date on the toxicity of imipenem versus
cefoxitin, our experience has been that imipenem is also more tolerable over a prolonged
treatment course. One nuance in using both antibiotics to treat MABc is that both
are dosed every 12 hours in contrast to the every 8-hour or every 6-hour schedule
on which they are dosed for most other infections. There does not appear to be utility
to continuous infusion dosing with cefoxitin and it is impractical with imipenem because
of drug stability.[72 ]
Intravenous Amikacin
IV amikacin is generally considered the most active antibiotic available for the treatment
of MABc infections.[1 ]
[5 ]
[36 ] Microbial killing due to IV amikacin is believed to be linked to peak serum-to-MIC
(C
max /MIC) ratios with the optimal ratio estimated as 3 to 5.[1 ]
[73 ] ATS treatment guidelines suggest a dose of 10 to 15 mg/kg, with the lower dose of
10 mg/kg recommended in the elderly or those on prolonged therapy (situations applying
to most patients).[1 ]
[36 ] The concern with IV amikacin is drug toxicity with ototoxicity/vestibular and renal
toxicity being the primary considerations, although liver function testing abnormalities
also occur. In general, ototoxicity seems to be irreversible and related to cumulative
dose and higher trough levels, while renal toxicity is usually reversible with drug
discontinuation/dose reduction and vestibular toxicity is often transient, regardless
of dosing change.[74 ]
[75 ]
[76 ]
[77 ] Toxicity seems to increase with treatment courses longer than 8 to 12 weeks.[77 ] It is controversial whether daily versus TIW dosing is more efficacious and there
are no good comparative data.[1 ] However, there are comparative data showing lower toxicity rates with TIW dosing,
so this is generally our recommended strategy. There does not appear to be any benefit
to dosing more than once a day as is done with many other bacterial infections, so
we strongly recommend against this for MABc.[76 ] Given the toxicity, close monitoring is important, and we recommend baseline and
monthly audiology testing as well as at least a complete metabolic panel and complete
blood count every other week as well confirming correct amikacin dose via estimation
of serum-to-MIC levels as above. In vitro data suggest that subinhibitory amikacin
levels may be not just less effective but also harmful by causing physiologic changes
to the organism that make it more robust and virulent.[78 ]
[79 ] Even more concerning, in vitro data suggest that MABc organisms with a functional
erm gene may be able to induce amikacin resistance on macrolide exposure through a WhiB7-dependent
network of resistance genes.[80 ] These data need additional confirmatory work and in vivo evaluation, but it is worrisome
given the prevalence of both macrolides and amikacin in many regimens and raises the
question of whether this contributes to the poor outcomes for subsp abscessus where a functional erm gene is frequently present and macrolides still often used.
Inhaled Amikacin
Since the most limiting feature of IV amikacin is drug toxicity, it is not surprising
that there has been interest in whether an inhaled formulation of that drug has efficacy
for treating MABc pulmonary infections. A retrospective study of 15 patients suggested
that inhaled amikacin led to improved symptoms and culture responses in a notable
minority of the cases, but that increasing toxicity limited doses higher than 250
mg/day.[81 ] Subsequent animal model work suggested a new liposomal form of inhaled amikacin
might be effective for the treatment of MABc (as well as MAC).[82 ] This led to one of the few randomized controlled trials in this field using inhaled
liposomal amikacin for NTM lung disease which showed improvement in the intervention
arm, although only one-third of the 89 patients were MABc and their response was difficult
to observe in the only 3-month-long study.[83 ] There is currently an ongoing open-label trial evaluating the liposomal formulation
in subsp abscessus lung infection (NCT03038178), which may shed more light on this issue.
Macrolides
Antibiotics in the macrolide class (clarithromycin, azithromycin, erythromycin) have
traditionally been the most important agents in the treatment of NTM infections. However,
as mentioned above, their use in the treatment of MABc infections is much more complicated
because of inducible macrolide resistance conferred by a functional erm gene. Selection of macrolide resistance is sometimes even less straightforward than
presence or absence of a functional erm gene as certain subsp abscessus sequevars and subsp bolletii appear to have other mechanisms that cause macrolide nonsusceptibility.[84 ] Although, traditionally, clarithromycin (CLARI) has been the macrolide of choice,
data suggest the newer macrolide azithromycin (AZI) is more active, easier to tolerate,
and causes somewhat less activation of the inducible macrolide resistance mechanism.[85 ] AZI has the added benefit of fewer drug–drug interactions as well as daily dosing
instead of twice-daily dosing with CLARI, and for these reasons, it should be considered
the macrolide of choice for MABc.[86 ] As mentioned, there is concern about potential antagonism by macrolides and amikacin
in MABc strains with an active erm gene, as well as with fluoroquinolones specifically in subsp abscessus (discussed in more detail in next section).[80 ] Interestingly, for subsp massiliense , there is possible in vitro synergy between macrolides (CLARI was macrolide studied)
and the fluoroquinolone moxifloxacin as well as the tetracycline tigecycliane, another
example of the different drug response phenotype of this subspecies.[87 ] One other difficulty created by macrolide and aminoglycoside combination therapy
is that macrolides can rarely cause sensorineural ototoxicity, which in a small number
of cases can be irreversible.[88 ] This effect can start after limited macrolide exposure and can make it difficult
to know whether it is the aminoglycoside or the macrolide causing ototoxicity.[88 ] The anti-inflammatory properties that macrolides, particularly AZI, possess have
been well documented as part of their benefit against airway exacerbations in many
different lung diseases states including non-CF bronchiectasis.[89 ]
[90 ] It is possible as part of a long-term regimen that they benefit to MABc pulmonary
infections in this way, but this has not been proven and, when used in this fashion,
they should not be counted as an active MABc antibiotics, unless the organism is known
to be macrolide susceptible.
Clofazimine
Clofazimine (CFZ) is an older antibiotic that has long been used in treatment of the
mycobacterial infections caused by Mycobacterium leprae and more recently included in some new regimens for drug-resistant tuberculosis.[91 ]
[92 ] CFZ MICs to MABc are typically very low and make it the most reliably active in
vitro oral antibiotic option available ([Table 3 ]).[36 ]
[93 ] In the context of MABc therapy, CFZ also has the attractive property of in vitro
synergy with amikacin, which seems to decrease MICs to both by four- to eightfold.[93 ]
[94 ]
[95 ] It also appears to have some synergy with macrolides.[95 ] These findings make CFZ a potentially useful agent in MABc therapy, and it is used
off-label frequently.[1 ] Two recent studies, one retrospective cohort (15 patients for initial therapy and
27 patients for salvage therapy) and one prospective observational cohort (54 patients),
used CFZ as part of a multidrug regimen for MABc pulmonary disease and showed promising
symptom and culture-based outcomes with relatively minimal rates of treatment-limiting
side effects.[60 ]
[61 ] Given these data, we recommend CFZ as an important part of many MABc treatment regimens,
especially when there are no other active oral agents. Although skin discoloration
is the most well-known of CFZ side effects, the most common treatment-limiting side
effects are gastrointestinal in nature.[60 ]
[61 ] CFZ has a very long half-life of roughly 50 to 75 days and a provider should assume
it takes around 2 months before the drug reaches therapeutically active levels.[96 ]
[97 ] CFZ has not been available through traditional pharmaceutical distribution since
2004 and can only been obtained either via investigational new drug (IND) directly
from the manufacturer (Novartis) or through individual IRB-approved IND to the Federal
Drug Administration (FDA). There is an ongoing placebo-controlled trial funded by
the FDA looking at the use of CFZ in MAC pulmonary disease, which should shed further
light on the drug's use in pulmonary NTM disease (NCT02968212).
Fluoroquinolones
Fluoroquinolones (FQs) are an attractive option for MABc therapy because of their
excellent oral bioavailability and good performance in the treatment of other mycobacterial
infections, especially tuberculosis. However, there are a couple of in vitro considerations
that somewhat mitigate this optimism for MABc as resistance to FQs has been shown
to develop quickly during monotherapy, and there is evidence of antagonism between
macrolides and FQs in many subsp abscessus isolates.[98 ]
[99 ]
[100 ] As previously mentioned, FQs are used in many other clinical diseases and ubiquitously
in agriculture, and this may contribute to the fact that high percentage of isolates
are FQ resistant, and thus the drug class should not be relied on without susceptibility
testing on a recent isolate.[101 ] That said, FQs have been used as part of combination therapy in a couple of retrospective
studies and there is probably a role for their use when MIC testing shows favorable
results.[9 ]
[102 ] Given the concerning in vitro studies with FQ monotherapy and potential antagonism
with macrolides for subsp abscessus (much more with CLARI than AZI), we strongly recommend against monotherapy or dual
therapy with only CLARI as part of a treatment or suppressive regimen.[98 ]
[99 ]
[100 ] When used for gram-positive infections such as mycobacterial infections, moxifloxacin
is generally the most active drug in the class and is the FQ we recommend if these
drugs are to be used for MABc therapy. Since therapy is often prolonged and the patient
population often receives intermittent steroids or other QT prolonging agents, care
should be taken to monitor for both QT prolongation and arthropathies when using FQs
for MABc therapy.
Tigecycline
The tetracycline (TCN) drug class has broad antibacterial activity including against
some mycobacteria. The newer IV TCN, tigecycline, appears to have particular in vitro
efficacy against rapid-growing mycobacteria including MABc compared with other TCNs
and to its own activity against slow-growing mycobacteria.[103 ]
[104 ] Here again, synergy with other drugs is an important consideration, and in vitro
data suggest that tigecycline may have synergy with macrolides, although it does not
appear to have this benefit with aminoglycosides.[105 ] Clinical data are limited, but one study that used tigecycline as part of a salvage
therapy regimen in 52 patients with MABc/M. chelonae infection (roughly 75% were MABc) showed clinical improvement in 62% of patients
who received more than 1 month of therapy out of the subset of 36 patients who had
clinically evaluated outcomes.[106 ] Adverse drug events were common, and although it is unclear which antibiotic in
the treatment was the cause of the symptoms, many of these events were gastrointestinal
in nature (nausea > vomiting > diarrhea > anorexia), side effects particularly common
with tigecycline.[106 ] The side effect profile was more pronounced at higher doses and mitigated by preinfusion
treatment with antiemetics; hence, in pulmonary disease we recommend dosing at 50
mg/day and using preinfusion antiemetics when the tigecycline is used.[106 ] When used in this manner, we recommend tigecycline as an alternative IV option if
either amikacin or imipenem/cefoxitin is not possible based on susceptibility testing
or other factors.
Oxazolidinones
The oxazolidinone drug class, which includes linezolid (LZD) and tedizolid (TZD),
acts by inhibiting protein synthesis at the 50s ribosomal subunit and has the attractive
properties of good oral bioavailability and tissue penetration.[107 ] LZD has been studied extensively for the treatment of tuberculosis and seems effective
and more tolerable at doses of either 300 or 600 mg daily compared with the 600 mg
twice daily dose used for nonmycobacterial infections.[108 ]
[109 ] Although the option of an active oral agent is appealing, clinical data for the
oxazolidinones in the treatment of MABc are at the level of case reports, and MICs
to a majority of MABc isolates are much higher than for TB, although good in vitro
activity can be seen in a minority of isolates.[1 ]
[110 ]
[111 ] LZD appears to have in vitro synergy with amikacin and macrolides in a proportion
of isolates, although it may have potential antagonism with both cefoxitin and moxifloxacin.[87 ]
[111 ] TZD may have better MABc activity than LZD with lower MICs on in vitro testing—however,
interpretive breakpoints do not yet exist for TZD.[110 ] Unfortunately, toxicity with LZD appears to be common in prolonged treatment courses
even at daily dosing and most individuals cannot tolerate treatment for more than
4 to 6 months.[112 ] Peripheral neuropathy, not mitigated by pyridoxine, is the most common side effect
seen in roughly 25% of cases.[112 ] In general, we do not recommend the oxazolidinones as part of first-line regimens
given the paucity of clinical data. However, they may have a role for periods of time
when supported by favorable susceptibility testing.
Bedaquiline
Bedaquiline is a novel diarylquinoline antibiotic that potently inhibits ATP synthase
in mycobacteria, has good bioavailability, and is now FDA approved for the treatment
of drug-resistant tuberculosis.[113 ]
[114 ] It appears to have a broad range of in vitro activity against NTM isolates including
members of the MABc.[115 ]
[116 ] Specifically, in this complex, roughly 70 to 80% of isolates appear to be susceptible
based on in vitro testing and estimated, but not formally established, breakpoints.[115 ] Data for NTM disease are limited, but safety data can probably be approximated from
the tuberculosis literature and suggest that nausea, hepatotoxicity, and QT prolongation
are the primary drug-related adverse events.[117 ] Clinical efficacy data are very limited, with the primary clinical study published
to date (a 10-patient case series: 6 MAC lung disease and 4 MABc lung disease) using
the 400 mg daily followed by 200 mg TIW dosing that has been used in the tuberculosis
trials.[118 ] This study suggested some microbiologic response based on semiquantitative sputum
cultures, although it also demonstrated high rates of nonsevere side effects.[118 ] Given these limited data, we recommend considering bedaquiline in alternative or
salvage therapy regimens when other oral antibiotic options are limited, although
ability to effectively utilize the medication will likely be constrained by cost and
drug access in some cases.
Specific Disease State Treatment Considerations
Specific Disease State Treatment Considerations
Cystic Fibrosis
CF is an important underlying predisposition for pulmonary NTM infection and disease
because of the significant bronchiectasis and impairment to pulmonary immune defenses
that CF confers. As a result, a high percentage of individuals with CF have pulmonary
colonization with an NTM, ranging from 5 to 30% depending on the case series.[18 ]
[26 ]
[119 ]
[120 ]
[121 ] MABc is the second most common NTM after MAC in this setting, but for reasons that
are unclear its incidence is rising, unlike MAC incidence, which is flat. Most NTMs
can cause invasive disease in CF, leading to more rapid decline in lung function and
at times to NTM-related mortality, with MABc seeming to contribute to a more deleterious
course than other NTMs.[121 ]
[122 ]
[123 ] Although there is concern for human transmission in MABc outbreaks at CF centers,
the limited genetic diversity in the complex makes genetic clustering harder to use
as an outbreak marker; therefore, while there is concern about this issue, it is generally
considered an unresolved concern.[25 ]
[124 ] Despite the importance of MABc in CF, there are no randomized controlled trials
or highly controlled studies looking at the effect of antimicrobial therapy versus
no therapy or comparing different antimicrobial regimens in CF, and treatment guidance
is limited to small, retrospective, single-center case series.[123 ]
One common practice in CF management, the use of prophylactic azithromycin to prevent
pulmonary exacerbations, does appear to be protective against the development of incident
NTM lung disease, although it is unclear if this applies to MABc, with some datasets
suggesting lower MABc prevalence and others suggesting the NTM prevalence shifting
toward MABc.[125 ]
[126 ] One concern given the antibiotic exposure in the CF population is that their MABc
isolates may in general be more resistant than in non-CF cases, but this has not been
clearly borne out by data.[127 ] Once MABc pulmonary disease develops, the same principles that apply to MABc treatment
in the non-CF population should be applied.[1 ] Similar to the non-CF setting, the subsp massiliense appears to be associated with better treatment outcomes than subsp abscessus .[128 ] Given the young age of this patient population, repeated treatment courses or long-term
suppression is more likely to be needed because of the many years of potential for
reinfection and/or relapse. In the CF population, there are even more issues with
concurrent pulmonary disease caused by other NTMs, especially MAC, as well as other
respiratory pathogens such as Staphylococcus aureus , Pseudomonas , and Aspergillus .[28 ]
[123 ] Unfortunately, although surgery is an attractive option for difficult MABc lung
disease cases in other patient populations, there may be additional risks to this
intervention in the CF leading to less favorable outcomes.[30 ] Another difficulty is that for many more advanced CF cases, the only management
strategy available is lung transplant, but as is discussed more in the subsequent
section, at many centers lung transplant is not offered when there is concomitant
MABc infection.
Treatment Considerations Specific to Lung Transplant Recipients
Lung transplant raises four unique and complicated considerations in relation to MABc
infection: (1) pretransplant management, (2) viability of lung transplant in the setting
of MABc lung disease, (3) risk of posttransplant MABc lung disease, and (4) management
of posttransplant MABc lung disease.
In general, pretransplant management is the same as general MABc management mentioned
above, although we favor the most aggressive possible antimicrobial therapy in the
immediate pretransplant period. One important consideration for MABc lung disease
is that many centers consider MABc infection a contraindication for lung transplant,
which precludes CF patients with end-stage disease from having a transplant, which
is the only curative management strategy for their CF disease.[129 ] This contraindication is largely based on expert opinion and the literature to support
this practice is limited.[130 ]
[131 ] In fact, several more recent small retrospective series suggest that outcomes may
not be worse in those with pretransplant pulmonary cultures showing MABc, although
there seems to be a higher rate of treatable posttransplant surgical site infections
with MABc in these hosts.[129 ]
[132 ]
[133 ]
[134 ]
Lung transplant is the solid organ transplant with the highest risk of posttransplant
NTM lung disease.[135 ]
[136 ] There does seem to be a significant rate of transient and clinically insignificant
colonization in the posttransplant period, so being post–lung transplant does not
inherently change the dynamic that not all positive pulmonary NTM cultures require
treatment and transient colonization does not seem to portend a worse posttransplant
outcome.[137 ]
[138 ] MABc appears to be the most common posttransplant cause of NTM lung disease with
a risk of infection highest in the first 3 years after transplant.[3 ]
[136 ] There seems to be a trend toward worse posttransplant survival in cases that develop
NTM lung disease. However, it is not clear if this trend is further worsened in those
whose lung disease is caused by members of MABc, and there is evidence suggesting
that posttransplant MABc infection can be successfully managed with prolonged aggressive
therapy.[3 ]
[129 ]
[132 ]
[133 ]
[134 ]
[136 ]
[139 ] Treatment is no different than in the nontransplant population, although regimen
choice may be more difficult and closer toxicity monitoring needed because of potential
drug interactions with transplant medications.[1 ] We recommend the same criteria for MABc disease in the posttransplant setting as
the nontransplant population but would initiate aggressive MABc therapy if a patient
meets criteria for MABc disease without any period of observation off therapy. We
recommend prolonged therapy along with multidrug IV induction therapy and would generally
err on the side of treating longer than the nontransplant population because of the
underlying immunodeficiencies. Finally, we would be more inclined to continue some
form of chronic suppressive therapy in this population if there is an oral regimen
with tolerable side effects and toxicities, although this is not needed in all cases.
Conclusion
The MABc organisms are increasingly recognized as a pulmonary pathogen that requires
treatment in some individuals. Our belief is that most infected individuals eventually
progress to disease, although natural history studies to definitively substantiate
this are lacking. The MABc poses significant and unique treatment challenges because
of its extensive drug resistance profile and propensity for persistence. The antibiotic
regimens used for this disease involve both antibiotics less commonly used by pulmonary
or infectious diseases practitioners and those that require prolonged use and careful
monitoring. Response to therapy is highly dependent on MABc subspecies, and even with
optimal therapy, some organisms in the MABc still have suboptimal antibiotic treatment
outcomes. In settings of suboptimal response, lung-resection surgery is a consideration
in management. While MABc disease is both more common and problematic in the unique
populations of CF and lung transplant recipients, the organism can still be managed
in these settings. Because of the complexity of this disease, expert consultation
is a recommended part of management. Additional research is desperately needed to
better understand disease pathogenesis, discover novel antibiotic therapies, and identify
optimal treatment strategies, with one of the most important challenges being the
identification whether in vitro susceptibility data predict clinical response.
