ABSTRACT
Mycobacterium abscessus is an extensively drug-resistant opportunistic pathogen that can cause chronic otomastoiditis. There are no evidence-based treatment regimens for this severe infection. We treated four children with M. abscessus otomastoiditis with a structured regimen of topical imipenem and tigecycline, intravenous imipenem and tigecycline, and oral clofazimine and azithromycin and adjunctive surgery. This structured approach led to cure, with 1 year of follow-up after treatment. Adverse events were frequent, mostly caused by tigecycline.
This Journal section presents a real, challenging case involving a multidrug-resistant organism. The case authors present the rationale for their therapeutic strategy and discuss the impact of mechanisms of resistance on clinical outcome. Two expert clinicians then provide a commentary on the case.
CASE PRESENTATIONS
Case 1.A 7-year-old boy with a history that included left-sided mastoidectomy presented with earache, otorrhea, and mild (20 dB) conductive hearing loss. Otoscopy revealed bilateral small perforations with mild otorrhea. Computed tomography (CT) showed bilateral contrast enhancing opacification of the mastoid cavity and the middle ears. On magnetic resonance imaging (MRI), there was an area of restricted diffusion in the mastoidectomy cavity on the left side, consistent with pus (Fig. 1, upper row). Mycobacterial cultures of pus from both ears yielded M. abscessus, and the drug susceptibilities (by microdilution [1]) are recorded in Table 1.
MRI images of two of the patients. (Upper row, case 1) MRI demonstrated bilateral contrast enhancing opacification of the mastoid cavity and the middle ears (arrows), consistent with otomastoiditis. There was an area of restricted diffusion in the mastoidectomy cavity on the left side as a result of pus (asterisk). (Bottom row, case 2) MRI demonstrated apex petrositis (circled) and a retropharyngeal abscess (asterisk) on the right side. On CT, osseous destruction of the petrous apex was seen (open arrow).
MIC data and interpretation of initial isolates of the four patientsa
Case 2.A 9-year-old girl presented with persistent right-sided otorrhea and right-sided submandibular lymphadenopathy. Otoscopy showed a central perforation and thickened middle ear mucosa, as well as otorrhea. Audiometry revealed a moderate right-sided conductive hearing loss (45 to 80 dB). CT showed opacification of the entire right middle ear and mastoid with osseous destruction of the petrous apex and cortical interruption of the mastoid bone. MRI revealed that this area was continuous with a retropharyngeal abscess (Fig. 1, bottom row). Bone and abscess fluid samples yielded M. abscessus, and treatment was initiated.
Case 3.A 15-year-old girl presented with left-sided otorrhea and hearing loss. Otoscopy showed unilateral otorrhea with a tympanostomy tube. A CT scan showed mucosal swelling in the left middle ear and mastoid. She underwent a combined-approach tympanoplasty for removal of inflamed mucosa. Because of persisting otorrhea, otorrhea fluid cultures were performed, yielding M. abscessus.
Case 4.A 7-year-old boy presented with right-sided otorrhea. Otoscopy showed a central tympanic membrane perforation, thickened middle ear mucosa, and otorrhea. CT and MR imaging showed contrast enhancing opacification of the right middle ear and the mastoid consistent with otomastoiditis without osseous destruction. Mycobacterial culture of otorrhea fluid yielded M. abscessus.
CHALLENGE QUESTION
What is the optimal antibiotic treatment regimen for otomastoiditis caused by amikacin-resistant Mycobacterium abscessus?
A. Oral macrolides plus one oral agent to prevent macrolide resistance.
B. One intravenous drug selected by in vitro susceptibility, plus a macrolide.
C. Two or more intravenous antibiotics, then two oral antibiotics, all based on in vitro susceptibility.
D. Option C, but with the addition of topical strategies, i.e., surgery and/or topical antibiotics.
E. There is no antibiotic treatment regimen with proven efficacy in this condition.
TREATMENT AND OUTCOMES
We have developed a guideline-compliant (2) antibiotic treatment strategy that combines systemic and local treatment consisting of a 3-month intensive phase with intravenous imipenem-cilastatin (60/60 mg/kg/day in 2 to 4 doses) and tigecycline (2.4 mg/kg/day in 2 doses), oral treatment with azithromycin (10 mg/kg once daily), and clofazimine (50 mg once daily; if weight is >40 kg, then 100 mg once daily), with topical imipenem-cilastatin (5/5 mg/ml; three drops three times daily) and tigecycline (5 mg/ml; three drops three times daily); topical treatment was halted when otorrhea stopped (mean duration, 6 weeks; range, 4 to 7 weeks). In the middle of the intensive phase, we performed combined-approach tympanoplasties to remove inflamed mucosa from the mastoid and the middle ear and rinsed the inner ear with 5/5 mg/ml imipenem-cilastatin and 5 mg/ml tigecycline solutions. After the intensive phase, an oral continuation phase of 3 months of oral azithromycin and clofazimine was given to complete a total of 6 months of antibiotic treatment. For intravenous access, all patients received a peripherally inserted central catheter (PICC). Solutions for topical application were prepared by the hospital pharmacy by reconstitution and dilution of powder formulations for intravenous administration with 0.9% sodium chloride under aseptic conditions. The stability of the solutions was limited to 48 (imipenem-cilastatin) and 72 h (tigecycline) at 2 to 8°C (3, 4).
We chose tigecycline and imipenem-cilastatin for topical therapy because of the concentration-dependent bactericidal effect of tigecycline against M. abscessus and the observation that β-lactam antibiotics (imipenem, because cefoxitin is not registered in the Netherlands) are the driving force in the recommended amikacin-β-lactam-macrolide regimen in pharmacodynamic models (5, 6). We did not administer amikacin, as isolates from all four patients proved to be amikacin resistant owing to the A1408G mutation in the rrs gene, likely related to prior topical aminoglycoside exposure. Clofazimine was given from the start because of its slow tissue accumulation and for its synergy with macrolides against M. abscessus (7, 8). We administered azithromycin despite the inducible macrolide resistance recorded for isolates of three of the four patients (Table 1), based on in vitro observations that clofazimine can prevent or delay the activation of erm(41)-mediated inducible macrolide resistance (8) and for lack of an alternative with proven efficacy.
Treatment and follow-up.The baseline characteristics, duration of treatment, and occurrence of adverse events in these four patients are presented in Table 2 and visualized in Fig. 2.
Baseline characteristics and treatment details
Timelines of antibiotic treatment and surgery. The x axis depicts the time (in months), and colored bars represent the drugs used. Asterisks next to the bars represent the moments when a drug was stopped because of toxicity or complications. The line with the S marking shows the timing of surgery. I/T, imipenem-tigecycline eardrops; I, imipenem; T, tigecycline; C, clofazimine; A, azithromycin.
i. Case 1.Despite granisetron administration, the boy suffered from nausea, vomiting, anorexia, and feeding difficulties requiring temporary total parenteral nutrition (TPN). Six weeks into antibiotic treatment, laboratory investigation revealed elevated transaminase levels, indicating hepatotoxicity. TPN and tigecycline were discontinued, resulting in a sharp decline of transaminase levels. Tigecycline was reintroduced in reduced dosage, but after 9 weeks of treatment, both intravenous antibiotics were discontinued because of hepatotoxicity, hypertriglyceridemia, and high cholesterol levels. Oral treatment with clofazimine, azithromycin, and minocycline was continued for 7 more months.
Five weeks into treatment, bilateral tympanoplasties were performed. Because of positive mycobacterial cultures of the surgical specimens, tympanoplasties were repeated 3 weeks later. Mycobacterial cultures performed during this procedure remained negative. After 9 months of treatment, there was mild bilateral conductive hearing loss, identical to the start of treatment (20 dB).
ii. Case 2.Nausea and vomiting were suppressed by metoclopramide. After 2 weeks, intravenous treatment was complicated by a venous thrombosis in the superior vena cava. The PICC was replaced and the thrombosis was successfully treated with heparin. After 4 weeks of treatment, tympanoplasty was performed. Mycobacterial cultures of surgical samples remained negative. Intravenous antibiotics were stopped after 10 instead of 12 weeks because of progressive elevated transaminase levels, arthropathy, and fatigue. Oral antibiotics were continued to a total treatment duration of 6 months. Right-sided hearing loss (40 dB) remained.
iii. Case 3.This girl completed the planned antibiotic treatment but experienced severe nausea and vomiting in the intensive phase, suppressed by granisetron and dexamethasone. Dexamethasone treatment led to stretch marks and weight gain. After 3 weeks of antibiotic treatment, a second tympanoplasty and mastoidectomy was performed. She suffers from persistent unilateral hearing loss (40 dB).
iv. Case 4.Because of severe nausea and vomiting, granisetron and dexamethasone were started. Dexamethasone use led to weight gain and hypocortisolism. Both disappeared after dexamethasone was stopped. One week into treatment, laboratory testing showed leukopenia, thrombopenia, and progressive elevation of serum transaminase levels, which prompted treatment interruption until laboratory results had normalized. Three weeks later, the boy developed toxic dermatitis with hives and an extensive rash. All antibiotics were stopped again and sequentially reintroduced. Azithromycin, the suspected cause, was not restarted. After 1 month of treatment, progressive swallowing problems led to discontinuation of clofazimine. Thereupon, we decided to prolong the intravenous treatment to 6 months. Six weeks into antibiotic treatment, a right-sided combined-approach tympanoplasty was performed. Cultures of surgical samples did not yield M. abscessus. After 4 months of treatment, progressive hypertriglyceridemia and hepatotoxicity led to treatment cessation. All laboratory abnormalities normalized within 1 month after cessation. Persistent unilateral hearing loss (20 dB) remains.
There was no treatment regimen with proven efficacy in M. abscessus otomastoiditis. Overall, combined systemic and topical antibiotic treatment with multiple rounds of surgery for otomastoiditis caused by M. abscessus yielded good results in these four patients. All four attained culture negativity while on treatment, and no recurrences have been noted after a mean of 14 months (range, 8 to 19 months; Table 2) of follow-up after treatment. This compares favorably to a larger series of 12 children with M. abscessus otomastoiditis; 2 (17%) experienced treatment failure after treatment with widely varying regimens and various numbers and extents of surgical interventions (9).
The frequency and severity of adverse events is problematic. Intravenous tigecycline seems to be the main culprit, causing severe nausea and vomiting, as previously reported (10). The major adverse event was the social impact of the treatment, i.e., missing school for days or weeks and several admissions to hospital for either adverse events or surgery.
ACKNOWLEDGMENT
No external financial support was obtained for this study.
FOOTNOTES
- Received 13 June 2019.
- Returned for modification 25 June 2019.
- Accepted 7 October 2019.
- Accepted manuscript posted online 14 October 2019.
For the commentary, see https://doi.org/10.1128/AAC.02023-19.
- Copyright © 2019 American Society for Microbiology.