Outcomes of Short-Term Tigecycline-Containing Regimens for Mycobacterium abscessus Pulmonary Disease

ABSTRACT

Short-term intravenous tigecycline therapy during a 1-month initial phase may improve early microbiological response in patients with Mycobacterium abscessus pulmonary disease (PD). However, short-term use of tigecycline did not improve the long-term culture conversion rate of M. abscessus PD. Further studies on the efficacy of prolonged intravenous tigecycline-containing regimens are needed.

INTRODUCTION

The burden of nontuberculous mycobacteria pulmonary disease (NTM-PD) is increasing (1, 2). Mycobacterium abscessus has emerged as one of the most important causes of NTM-PD (3). Of M. abscessus subspecies, M. abscessus subspecies abscessus (here, M. abscessus) PD is the most difficult disease to cure (4 to 7). The major challenge in treating M. abscessus PD is its high rate of antibiotic resistance (8). Most M. abscessus strains have inducible resistance to macrolides related to the functional erm(41) gene (9). Guidelines recommend long-term multidrug therapy, including intravenous amikacin, imipenem (cefoxitin), or tigecycline (3, 10, 11). However, the outcomes are poor (5, 6), and long-term antibiotic use is not feasible due to related toxicity (12).

Tigecycline is a semisynthetic tetracycline (13). Studies have reported high in vitro activities of tigecycline against NTM isolates (14, 15). However, limited studies have reported on the efficacy of tigecycline for M. abscessus PD (16 to 19). Thus, we aimed to evaluate the microbiological response to short-term tigecycline-containing regimens versus conventional multidrug regimens without tigecycline in M. abscessus PD patients.

We investigated M. abscessus PD patients who received multidrug regimens between April 2013 and August 2020 at Samsung Medical Center, South Korea. Among patients treated for 12 months, we compared microbiological response within 12 months after treatment between groups who received regimens, including intravenous tigecycline in the initial phase (n = 28) or not (n = 36) (Fig. S1 in the supplemental material). Microbiological response, including acid-fast bacilli (AFB) culture negativity or culture conversion, defined as three consecutive negative tests, was analyzed. All data were obtained from an institutional review board-approved observational cohort (ClinicalTrials.gov registration no. NCT00970801). Data were retrospectively analyzed.

All patients with M. abscessus PD were treated with conventional agents, including amikacin (10 to 20 mg/kg/day) (or nonliposomal inhalation), imipenem (750 mg three times/day) (or cefoxitin), macrolide (250 mg/day), clofazimine (50 to 100 mg/day), linezolid, or rifabutin, while the tigecycline group additionally received intravenous tigecycline for 2 or 4 weeks during the initial phase of treatment. The dosage of tigecycline was stared with 25 mg/day. In the absence of adverse effects, the dose was titrated up to 50 mg/day, while the dosage was reduced when adverse effects, such as gastrointestinal symptoms, occurred. Detailed treatment modalities are shown in Table S1.

AFB smear and culture specimens were performed 1, 3, and 6 months after treatment initiation and then at 2- to 3-month intervals. In patients with persistently positive cultures, drug susceptibility testing (DST) was repeated every 6 months. DST was performed by measuring MICs using the broth microdilution method (19). The Pearson χ² test/Fisher exact test and Mann-Whitney U were used to compare categorical and continuous variables, respectively. Cox-proportional hazard analysis was used to evaluate factors associated with culture conversion within 12 months (R version 3.6.1, R Foundation for Statistical Computing, Vienna, Austria).

Characteristics of the study patients are shown in Table 1. Of all patients, the median age was 59 years, and most (80%) were female. The most common underlying condition was a previous tuberculosis, and 89% (57/64) had nodular bronchiectatic disease. Of them, 28 patients received intravenous tigecycline for the initial phase, and the remaining 36 patients did not receive tigecycline. There were no significant differences in characteristics between the two groups.

TABLE 1.

Characteristics of study patients^a

Characteristics	Tigecycline group (n = 28)	Nontigecycline group (n = 36)	P value
Age ≥65 yrs	9 (32)	10 (28)	0.918
Male	6 (21)	7 (19)	>0.999
BMI <18.5 kg/m2	9 (32)	13 (36)	0.947
Current smoker or ex-smoker	4 (14)	7 (19)	0.835
Underlying conditions
Previous tuberculosis treatment	12 (43)	21 (58)	0.329
Previous NTM treatment	12 (43)	11 (31)	0.450
Obstructive pulmonary disease	3 (11)	3 (8)	>0.999
Chronic pulmonary aspergillosis	4 (14)	5 (14)	>0.999
Symptoms
Cough	26 (93)	28 (78)	0.193
Sputum	26 (93)	31 (86)	0.650
Hemoptysis	7 (25)	9 (25)	>0.999
Radiologic form
Fibrocavitary disease	5 (18)	2 (6)	0.246
Nodular bronchiectatic disease	23 (82)	34 (94)	0.246
With cavity	11/23 (31)	11/34 (39)	0.642

^aData are presented as number (percentage) or median (interquartile range). BMI, body mass index; NTM, nontuberculous mycobacteria.

Detailed treatment modalities are shown in Table S1. All patients were administered intravenous amikacin, imipenem/cefoxitin, and oral macrolide or clofazimine in the initial phase for about 4 weeks. In the tigecycline group, tigecycline was added in the initial phase (median 28 days; interquartile range 24 to 28 days). In the continuation phase, oral macrolide, clofazimine, or inhaled amikacin (90% in tigecycline groups versus 56% in nontigecycline group, P = 0.008) was used. Four of the nontigecycline group received surgical lung resection within 4 months after antibiotics.

Microbiological responses were evaluated based on sputum AFB culture negativity and negative culture conversion within 12 months after treatment (Table 2). AFB culture negativity rate at 1 month was significantly higher in the tigecycline group than the nontigecycline group (89% [25/28] versus 50% [18/36], respectively, P = 0.002). Even after excluding the four surgically treated patients, AFB culture negativity rate at 1 month was still high in the tigecycline group (89% [25/28] versus 50% [16/32], P = 0.003). However, during the continuation phase, the rate of culture negativity decreased in both groups. The nontigecycline group tended to have more culture conversion within 12 months than the tigecycline-group (44% versus 26%, P = 0.213). However, there was no difference in microbiological responses, except culture negativity rate at 1 month. In multivariable analysis, initial tigecycline was not associated with culture conversion within 12 months. During the study period, one patient in the tigecycline group acquired resistance to macrolide (Table 3), while five acquired resistance to macrolide in nontigecycline-group.

TABLE 2.

Treatment outcome within 12 months after treatment^a

Outcome	Tigecycline group (n = 28)	Nontigecycline group (n = 36)	P value
AFB culture negativity
At 1 mo of treatment	25 (89)	18 (50)	0.002
At 3 mo of treatment	12 (43)	13 (36)	0.771
At 6 mo of treatment	9 (32)	15 (42)	0.603
At 12 mo of treatment	7 (26)	14 (39)	0.418
Culture conversion within 12 mo	7 (26)	16 (44)	0.213
Time to conversion, mo	0.6 (0.5 to 3.1)	2.2 (0.9 to 8.2)	0.169

^aData are presented as number (percentage) or median (interquartile range). AFB, acid-fast bacilli.

TABLE 3.

Changes in drug susceptibility testing within 12 months after treatment^a

Tigecycline group (n = 28)
Drug	Before treatment (n = 27)			After treatment (n = 20)
	Susceptible	Intermediate	Resistant	Susceptible	Intermediate	Resistant
Amikacin	26 (96)	1 (4)	0 (0)	17 (85)	2 (10)	1b (5)
	(≤16 μg/mL)	(32 μg/mL)	(≥64 μg/mL)	(≤16 μg/mL)	(32 μg/mL)	(≥64 μg/mL)
Cefoxitin	9 (33)	18 (67)	0 (0)	4 (20)	16 (80)	0 (0)
	(≤16 μg/mL)	(32 to 64 μg/mL)	(≥128 μg/mL)	(≤16 μg/mL)	(32 to 64 μg/mL)	(≥128 μg/mL)
Imipenem	17 (63)	10 (37)	0 (0)	9 (45)	11 (55)	0 (0)
	(≤4 μg/mL)	(8 to 16 μg/mL)	(≥32 μg/mL)	(≤4 μg/mL)	(8 to 16 μg/mL)	(≥32 μg/mL)
Clarithromycin	1 (4)	0 (0)	26c (96)	0 (0)	0 (0)	20c (100)
	(≤2 μg/mL)	(4 μg/mL)	(≥8 μg/mL)	(≤2 μg/mL)	(4 μg/mL)	(≥8 μg/mL)
Nontigecycline group (n = 36)
Drug	Before treatment (n = 35)			After treatment (n = 33)
	Susceptible	Intermediate	Resistant	Susceptible	Intermediate	Resistant
Amikacin	27 (77)	8 (23)	0 (0)	289 (85)	5 (15)	0 (0)
	(≤16 μg/mL)	(32 μg/mL)	(≥64 μg/mL)	(≤16 μg/mL)	(32 μg/mL)	(≥64 μg/mL)
Cefoxitin	9 (26)	25 (71)	1 (3)	4 (12)	29 (88)	0 (0)
	(≤16 μg/mL)	(32 to 64 μg/mL)	(≥128 μg/mL)	(≤16 μg/mL)	(32 to 64 μg/mL)	(≥128 μg/mL)
Imipenem	13 (37)	20 (57)	2 (6)	14 (42)	19 (58)	0 (0)
	(≤4 μg/mL)	(8 to 16 μg/mL)	(≥32 μg/mL)	(≤4 μg/mL)	(8 to 16 μg/mL)	(≥32 μg/mL)
Clarithromycin	4 (11)	0 (0)	31d (89)	2 (6)	0 (0)	31d (94)
	(≤2 μg/mL)	(4 μg/mL)	(≥8 μg/mL)	(≤2 μg/mL)	(4 μg/mL)	(≥8 μg/mL)

^aData are presented as number (percentage) or median (interquartile range).

^bIn the tigecycline group, one patient with cavitary nodular bronchiectatic disease was treated with multiple intravenous agents (amikacin, imipenem, tigecycline) and oral agents (azithromycin, clofazimine, and linezolid). However, the patient acquired resistance to amikacin after 8 months of treatment.

^cBefore treatment, 25 patients showed inducible resistance to clarithromycin, and 1 patient showed acquired resistance to clarithromycin, but after treatment, 18 patients showed inducible resistance to clarithromycin and 2 patients showed acquired resistance to clarithromycin.

^dBefore treatment, 30 patients showed inducible resistance to clarithromycin and 1 patient showed acquired resistance to clarithromycin, but after treatment, 25 patients showed inducible resistance to clarithromycin and 6 patients showed acquired resistance to clarithromycin.

In this study, AFB culture negativity rate at 1 month of treatment in the tigecycline group was high, up to 89%, but the rate rapidly decreased after the injections were stopped. This suggests that the use of tigecycline in addition to amikacin or imipenem/cefoxitin can help to reduce bacterial burden, but short-term addition of tigecycline may not improve long-term outcome. As in our previous data, we used tigecycline only for a short period (19), because tigecycline commonly caused severe nausea. Recently, in our institution, we have tried to extend the initial phase with tigecycline for patients who tolerate tigecycline well. However, there are concerns regarding adverse effects or dysbiosis associated with long-term use of broad-spectrum antibiotics. Another challenge is that tigecycline cannot be administered orally and an implantable catheter is necessary. Thus, complications such as vascular damage or infection are inevitable. Recent studies reported that omadacycline, an orally administrable tetracycline-class agent, had low MIC against M. abscessus, and a case of its use for skin infections was reported (20 to 22).

In the nontigecycline group, four patients underwent lung resection, and the conversion rate of the group within 12 months was relatively high. Presumably, this may be caused by the surgery playing a role in removing the bacterial source, which is difficult to achieve only by antibiotics.

Our study had several limitations. First, administration of tigecycline was made by the attending physician, and this may have biased the results. Second, because patients in the tigecycline group were more likely to have any cavitary lesions than those in the nontigecycline group, this may have influenced long-term culture conversion.

In conclusion, short-term use of tigecycline may improve early microbiological response in M. abscessus PD patients but did not improve the long-term culture conversion rate.