Published ahead of print on March 2, 2006, doi:10.1164/rccm.200509-1531OC
© 2006 American Thoracic Society doi: 10.1164/rccm.200509-1531OC
Factors Related to Response to Intermittent Treatment of Mycobacterium avium Complex Lung DiseaseDivision of Pulmonary and Critical Care Medicine, University of California, San Diego, San Diego; Division of Pulmonary and Critical Care Medicine, Stanford University Medical Center, Stanford, California; Department of Medicine, University of Texas Health Center, Tyler, Texas; Division of Pulmonary and Critical Care Medicine and Internal Medicine, Mayo Clinic, Rochester, Minnesota; Department of Medicine, New York University Medical Center/Bellevue Hospital, New York City, New York; and Division of Mycobacterial and Respiratory Infections, National Jewish Medical and Research Center, Denver, Colorado Correspondence and requests for reprints should be addressed to Antonino Catanzaro, M.D., Professor of Medicine, University of California, San Diego, 200 West Arbor Drive 8374, San Diego, CA 92103-8374. E-mail: acatanzaro{at}ucsd.edu
Rationale: Mycobacterium avium complex pulmonary disease (MAC-PD) is associated with substantial morbidity, and standard daily multidrug therapy is difficult to tolerate. Objectives: To characterize response to a three-times-weekly (TIW) regimen of clarithromycin, ethambutol, and rifampin.
Methods: A 1-yr prospective noncomparative trial of TIW treatment was conducted during 2000–2003 in 17 U.S. cities. Participants were 91 HIV-negative adults, diagnosed with moderate to severe MAC-PD, who originally participated in a trial of an inhaled IFN- Results: Treatment response rates (and median response times) were 44% (2 mo or longer) for culture, 60% (5.5–11.5 mo) for HRCT, and 53% (8.5 mo) for symptoms. Having noncavitary, compared with cavitary, disease increased culture response by 4.0 times (95% confidence interval [CI], 1.7–9.2) and HRCT response by 4.9 times (95% CI, 1.9–13.0). Culture response was 1.5 times (95% CI, 1.1–2.2) higher for older subjects and 2.2 times (95% CI, 1.0–4.7) higher for previously untreated subjects. Being smear-negative increased culture response by 2.3 times (95% CI, 1.1–5.2) but decreased HRCT response by 4.4 times (95% CI, 1.7–11.5). Increasing ethambutol use by 5 mo increased culture response by 1.5 times (95% CI, 1.0–2.1) but decreased symptom response. Not having chronic obstructive pulmonary disease, bronchiectasis, or poor lung function increased symptom response by 1.9 to 3.9 times. Conclusions: TIW therapy was less effective for MAC-PD patients with cavitary disease and a history of chronic obstructive pulmonary disease, bronchiectasis, or previous treatment for MAC-PD. Further research is needed to study the long-term outcomes of TIW treatment.
Key Words: clarithromycin • ethambutol • Mycobacterium avium complex • rifampin • risk factors Mycobacterium avium complex (MAC) refers to a group of ubiquitous environmental mycobacteria, M. avium and M. intracellulare. MAC is a pathogen that can produce a wide range of disease: disseminated MAC disease (DMAC) in immunocompromised individuals with acquired immunodeficiency syndrome and MAC pulmonary disease (MAC-PD) in immunocompetent adults (1). MAC disease is associated with substantial morbidity and mortality. Although DMAC rates declined dramatically after 1995 with the introduction of highly active antiretroviral therapy (2), pulmonary physicians observed an increase in MAC-PD among HIV-negative patients in the late 1990s and early 2000s (2, 3). The optimal treatment regimen for MAC-PD has not been fully established (4, 5). The most clinically efficacious drugs for the treatment of MAC are the new macrolide or azalide antibiotics, clarithromycin or azithromycin, respectively. These drugs are generally part of a multidrug regimen that includes rifamycin and ethambutol (4, 6). Before the availability of rifabutin, rifampin was the most commonly used rifamycin to treat MAC-PD. The standard MAC-PD treatment typically involves a three- or four-drug (with or without an aminoglycoside) regimen administered daily up to 24 mo (4). This regimen becomes expensive, is poorly tolerated, and can lead to a variety of side effects (e.g., ocular toxicity, gastrointestinal events, and abnormal body chemistry) (7, 8). Recently, results of two novel studies have suggested that a three-times-weekly (TIW) regimen including clarithromycin or azithromycin may be as effective as a daily regimen (9, 10). This TIW regimen holds promise in reducing cost, side effects, and premature treatment discontinuation (4, 10). The current study provides a unique opportunity to prospectively examine the effects of a TIW regimen in a relatively sizeable sample of patients with well-characterized MAC-PD from across the United States. The aims of the current analysis are (1) to evaluate bacteriologic, radiographic, and symptomatic response rates and response times and (2) to determine key factors that are associated with beneficial response to TIW treatment.
Study Design and Data Source This study uses data collected as part of a randomized, double-blind, placebo-controlled trial of standard triple-drug TIW therapy with or without inhaled IFN-  for the treatment of MAC-PD. The duration of therapy was originally 52 wk and was extended to 72 wk in midstudy to allow for longer time to observe culture conversion.
Interim analysis of trial data revealed no difference in outcomes between IFN-
In addition to inhaled IFN-
Participants
Procedures
Clinical assessment.
Radiographic assessment.
Bacteriologic assessment.
Main Outcome Variables
Culture improvement was defined, based on the definition from Griffith and colleagues (10), as a reduction in colony count in three consecutive sputum cultures spanning at least 60 d and subsequently sustained through Week 52 or the end of the observation period. Specifically, culture improvement referred to any one of the following situations: (1) conversion from 4+ or 3+ to 1+, countable colonies, or negative (i.e., from Study radiologists determined the presence of five radiographic features (i.e., cavitary disease, bronchiectasis, nodules, consolidation, and tree-in-bud appearance) at baseline and classified each feature as improved, unchanged, or worsened during follow-up. HRCT improvement was defined as improving in at least one radiographic feature at any time, lasting until the end of the observation period. In determining HRCT improvement, change in cavitary disease was given the most weight and tree-in-bud the least weight; bronchiectasis, nodules, and consolidation were given equal weight. The symptoms under study were cough, sputum production, hemoptysis, fever, chills, night sweat, and dypsnea, all of which were given equal weight. Symptom improvement was defined as improving in at least one of the seven symptoms for at least 60 d and subsequently sustained through Week 52 or the end of the observation period.
Statistical Analysis Because outcomes were right-censored for some study subjects, mainly due to early study termination, survival analysis was conducted. Rates of culture, HRCT, and symptom improvement were calculated for the total study group and by cavitary and noncavitary subgroups. Time to culture, HRCT, and symptom response were presented as Kaplan-Meier survival curves, each stratified by cavitary and noncavitary subgroups. The subgroup curves were compared by using the log rank test, and median survival times were calculated.
Separate Cox proportional hazards models were used to identify key factors associated with culture, HRCT, and symptom response to TIW treatment. Hazard ratios and 95% confidence intervals were used to report relative response rates among groups. A best-subset selection procedure helped to produce parsimonious models from a set of factors previously judged to be of clinical importance (13). High global score
Tests of significance were two-sided at an
Participants Ninety-one patients were treated for MAC-PD with a multidrug regimen administered TIW. Sixty-four percent (58/91) of the subjects completed the study (observation time  343 d). The other subjects were in the study for a shorter duration due to adverse events (n = 8), early withdrawal by patients (n = 7), and early study discontinuation prompted by investigators/sponsor (n = 18: six subjects were never culture positive for MAC after the initial screening, two were noncompliant, six developed resistance to clarithromycin, and four enrolled close to the end of the study). In this study, 19 subjects had cavitary disease without bronchiectasis, 30 had both cavitary disease and bronchiectasis, 39 had bronchiectasis without cavitary disease, and 3 had neither. In univariate analysis, the cavitary disease group (63% with bronchiectasis) and the noncavitary disease group (93% with bronchiectasis) differed significantly on a variety of factors at baseline (Table 1). Participants with noncavitary disease tended to be older on average (68 vs. 61 yr in the cavitary group) and female (74 vs. 45%). Those with cavitary disease were more likely to be previous or current tobacco smokers (82 vs. 38% in the noncavitary group) with longer duration of tobacco use (27 vs. 10 yr), to have a history of COPD (65 vs. 33%) or TB (23 vs. 2%), to be AFB smear positive at baseline (62 vs. 30%), to have undergone MAC treatment before the start of the study (59 vs. 47.7%), and to have severe dyspnea (63 vs. 33%) or consistent sputum production (46 vs. 24%) at baseline. The groups did not differ in performance on spirometry, although the lower FEV1/FVC ratio was consistent with increased COPD in the cavitary group.
Treatment Dose and Duration The numbers of patients treated with each drug were as follows: 80 with clarithromycin, 9 with azithromycin, 2 with both clarithromycin and azithromycin, 90 with ethambutol, 86 with rifampin, and 4 with rifabutin. One subject received azithromycin without ethambutol or a rifamycin but was included in the intent-to-treat analysis; excluding this individual did not change the reported results. At screening, 94% of subjects harbored mycobacteria that were susceptible to clarithromycin at a concentration of 4 µg/ml; 4% were susceptible at 8 µg/ml, and 2% had intermediate resistance at 16 to 32 µg/ml. The average dose and duration of use were 911 mg and 265 d for clarithromycin, 636 mg and 343 d for azithromycin, 1,402 mg and 277 d for ethambutol, 569 mg and 277 d for rifampin, and 450 mg and 198 d for rifabutin. Overall, the average drug dose did not differ significantly among groups who did and did not improve on culture, HRCT, and symptoms. Duration of drug use tended to be longer in the group who responded, especially by culture, but the difference between groups was also not statistically significant (data not shown).
Treatment Response Rate and Time
When culture conversion was defined as conversion from a positive to negative culture (i.e., to no growth), the response rate was approximately 13% (4% in the cavitary disease subgroup and 24% in the non–cavitary disease subgroup, p = 0.010). Of these 12 subjects with culture conversion to negative, 83% (n = 10) were women, 100% had a history of bronchiectasis, and 42% (n = 5) were previously treated for MAC-PD for at least 6 mo. Among participants who were culture positive for MAC, the median time to culture improvement was 67 d from baseline for the noncavitary subgroup; the cavitary subgroup never achieved the 50% response level (p value < 0.001; Figure 2A). Among those with HRCT evaluations, the median time to radiographic improvement was 164 d for the noncavitary group and 340 d for the cavitary group (p value = 0.085; Figure 2B). Among those with symptom assessment, the median time to symptom improvement was 252 d for both noncavitary and cavitary subgroups (p value = 0.806; Figure 2C).
Factors Associated with Treatment Response Clarithromycin levels are decreased by both rifampin and rifabutin, although to a greater extent by rifampin (14). Groups treated with the combinations of clarithromycin/rifamycin (rifampin or rifabutin) and azithromycin/rifamycin did not differ significantly in improvement rates. This may simply be due to the small sample size, particularly in the azithromycin group, but outcome differences cannot be properly assessed from these data. Compared with the clarithromycin group, the azithromycin group had slightly higher improvement rates by culture (55 vs. 43%, p = 0.526) and HRCT (71 vs. 59%, p = 0.687) and comparable symptom improvement rate (55 vs. 52%, p = 1.000). There were also no significant differences in treatment response parameters between clarithromycin/rifampin and azithromycin/rifampin combinations.
Table 2 shows the key factors associated with improvement in culture, HRCT, and symptom assessments. Culture response was significantly related to disease subgroup at baseline, AFB smear result at baseline, history of previous treatment for MAC, age, and duration of ethambutol use (p
HRCT response was significantly related to baseline disease subgroup and AFB smear results (p  0.002). HRCT response rates were 4.93 (95% CI, 1.88–12.96) times higher for participants who had noncavitary disease (compared with those with cavitary disease) and 4.41 (95% CI, 1.69–11.54) times higher for those who were AFB smear positive (compared with those who were smear negative; Table 2). HRCT response rates tended to be higher for participants with lower FEV1 (by 1 L), history of TB, or no history of COPD, although these associations were not statistically significant.
Symptom response was significantly related to history of COPD, history of bronchiectasis, and FEV1, and FVC (p
Adverse Events
This study is the largest prospective evaluation of the treatment of MAC-PD using an intermittently administered regimen, and findings from the study have helped to identify the types of patients who may benefit from the TIW regimen. Culture and HRCT response rates, adjusted for confounders, were four to five times higher in patients with noncavitary than cavitary disease. Older participants and those who were not previously treated for MAC had about twice the culture response rates of those 10 yr younger or those previously treated. Participants who were AFB smear negative at baseline were twice as likely to have culture response but four times less likely to have HRCT response. Longer duration of ethambutol use increased culture response rate by 1.4 times but equally decreased the symptom response rate. Participants who did not have a history of COPD or bronchiectasis, had lower FEV1, or had higher FVC had two to four times increased rates of improvement in symptoms. Although the majority of participants experienced some side effects, approximately 25% had serious adverse events.
One of the limitations of the current study is that the original study was designed to test the effectiveness of IFN- Another potential limitation is that the optimal effects of TIW treatment may not be clear due to inclusion of previously treated subjects. However, the study reflects the moderate to severe MAC-PD population seen in the clinical setting, which consists of many patients who have undergone multiple cycles of therapy. Previous treatment is a relevant issue because it is known to be negatively associated with treatment response (10, 15); in our study, this association (p < 0.05) was independent of demographic characteristics, risk factors, medical history, lung function, and symptoms. Another potential limitation of the study is incomplete data for subjects without HRCT readings and for those who did not complete the study. The subjects with incomplete data were comparable to subjects with complete data in factors evaluated in the study, such as demographic characteristics and history of bronchiectasis or cavitary disease, and in study outcomes; an exception was for symptom improvement, which occurred significantly more frequently in the group that completed the study evaluations. Therefore, except for symptom outcome, the study findings are unlikely to be biased due to incomplete data. The incomplete participation rate (16% patient-initiated and 36% overall) is also comparable to the rates observed from previous studies on TIW treatment (20–30%) (10) and other studies on macrolide-containing regimens (11–33%) (16). Patients' poor tolerance to MAC medications is a known reason for discontinuation of treatment. Therefore, it is important to include study subjects who did not complete the study to obtain results that are more generalizable to the clinical environment. There are several other limitations in this study. The study duration of 52 wk is arbitrary and may not allow sufficient time to observe significant treatment effects. Because directly observed therapy was not administered, nonadherence to the treatment regimen may have caused the low response rates in this study. Although improvements in culture, HRCT, and symptoms were attributed to TIW treatment, it is possible that some improvements could have resulted even without treatment (4).
In this study, the response rates were derived from follow-up on all study participants up to loss to follow-up or study termination. The culture response rate was comparable to rates of previous studies if intent-to-treat analyses were compared; to date, the overall success rate of macrolide-containing regimens in previous studies has been 56%, ranging from 26 to 71%, using intent-to-treat estimates (16). It is important to note the distinction between culture improvement and infection resolution. In this study, culture response was defined as substantial reduction in MAC colony count, a definition used in established literature (10) but which has unclear clinical significance. The clinical definition of culture conversion to negative (i.e., no growth) yielded a substantially lower response rate compared with the range from previous studies. The low response rates cannot be attributed to the inhaled IFN- Our study results show that having cavitary or noncavitary disease drives the difference in clinical presentation (Table 1) as well as response to TIW treatment (Table 2). For example, culture response occurred in 21% of patients with cavitary disease but without bronchiectasis, in 20% of patients with both cavitary disease and bronchiectasis, in 72% of patients with bronchiectasis but without cavitary disease, and in 67% of patients with neither bronchiectasis nor cavitary disease. Similarly, HRCT improvement occurred in 40% of patients with cavitary disease but without bronchiectasis, in 50% of patients with both cavitary disease and bronchiectasis, in 75% of patients with bronchiectasis but without cavitary disease, and in 100% of patients with neither bronchiectasis nor cavitary disease. Because presentation and outcomes were mainly different between the cavitary disease group (with or without bronchiectasis) and the non–cavitary disease group (predominantly bronchiectasis), the categorization of cavitary/noncavitary disease was appropriate in this study. The culture response rate for participants with noncavitary disease in this study (71%) was comparable to estimates from other TIW and daily-treatment studies (e.g., 78 and 82% [10, 15]). However, the response rate for those with cavitary disease in this study (20%) was substantially lower than estimates from other studies (e.g., 63 and 74% [10, 15]), although these estimates from previous studies were reported only for subjects who completed the studies. The low overall response rate in this study cannot be fully accounted for by the proportion of participants with cavitary disease; 54% of subjects in this study had cavitary disease, compared with 41 and 46% in other studies (10, 15). A partial explanation may be the fact that the patients with moderate or severe MAC-PD in the current study had greater pulmonary comorbidities and thus were less likely to respond to TIW treatment. In general, patients with preexisting lung diseases tend to have slower sputum culture conversion (3). In some previous studies, 33 to 48% of participants had other underlying pulmonary conditions (10, 15). By comparison, nearly all patients with MAC-PD in this study had a history of other lung diseases; 70% (57% of women and 90% of men) had a history of COPD, TB, or recurrent pneumonia.
Overall culture response rate (44%) was lower than HRCT response rate (60.4%) and symptom response rate (52.5%). Although the study radiologists were blinded to subjects' treatment group (i.e., IFN- Despite the difference in culture response rates across studies, factors that determined response were similar. In this study, culture response rates were lower for participants who had cavitary disease at baseline, who were AFB smear positive at baseline, or who were persistent cases treated previously for MAC but who remained susceptible to clarithromycin. In addition, the current analysis showed that use of ethambutol for 5 mo longer may increase the response rate by 45%; ethambutol use was correlated to the duration of the triple-drug therapy. Other distinctive contributions of this study involved evaluating multiple dimensions of MAC-PD beyond culture response, the main focus of most previous studies. The current analysis identified factors that determined HRCT response (factors such as having noncavitary disease or being AFB smear positive at baseline) and symptom response (factors such as not having a history of COPD or bronchiectasis or having higher FVC or lower FEV1) to TIW treatment. Generally, we found that subjects with relatively moderate illness were more likely to improve with treatment. Yet, in some instances, more severe presentations may allow greater margin for improvement. For example, subjects with lower FEV1 may have greater degree of airway inflammation, which may be more amenable to improvement in symptoms with antiinfective and antiinflammatory treatment. In addition, this study identified median times to treatment response (e.g., about 2 mo for culture response, 5.5 mo for HRCT response, and 8.5 mo for symptom response in the non–cavitary disease group; about 11.5 mo for HRCT response and 8.5 mo for symptom response in the cavitary disease group). Furthermore, the frequency of side effects was equivalent to results from previous studies on daily treatment; however, whether or not the events were due to the triple-drug therapy remains unclear in this study. Overall, results from this study suggest that TIW therapy may not be beneficial for all patients with MAC-PD. On the basis of culture, HRCT, and symptom findings, this treatment strategy seemed less effective in patients with cavitary disease, particularly if they had a history of COPD, bronchiectasis, or previous treatment for MAC-PD. Further evaluation of long-term effects of TIW therapy on MAC-PD progression remains important. The selection of patients for TIW versus daily administration of therapy requires a careful evaluation of the risks and benefits of each treatment strategy and the overall goals of therapy. Assistance from experts in the management of MAC-PD is recommended when treatment decisions and endpoints are not well defined.
The authors thank Dr. Sima Faris-Young and the staff at InterMune, Inc., for coordination of the clinical trial on inhaled IFN- treatment, and the study coinvestigators (Drs. Nancy Dunlap, Gwen Huitt, Mark Gotfried, Aaron Milstone, Kenneth Olivier, Juzar Ali, Gerard Criner, Sigredo Aldarondo, Brian Morrissey, Richard Duma, and James Cook) for their crucial roles in carrying out the study at their institutions.
Originally Published in Press as DOI: 10.1164/rccm.200509-1531OC on March 2, 2006
Conflict of Interest Statement: P.K.L. does not have a financial relationship with a commercial entity that has an interest in the subject of the manuscript. D.E.G. does not have a financial relationship with a commercial entity that has an interest in the subject of the manuscript. T.R.A. has received a research grant in 2001–2003 from InterMune for participating in a multicenter study on inhaled IFN- Received in original form September 29, 2005; accepted in final form March 1, 2006
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
2 statistics were used to select the best predictor variables within each model size. To select the best model size from nested models, likelihood ratio tests were used to successively delete nonsignificant variables, unless the removal of the variable changed the hazard ratios of remaining factors by 15% or more. 
level of 0.05, and p values less than 0.05 were considered significant. Best-subset selection was performed using SAS statistical software (SAS 9.1 for Windows, 2002–2003; SAS Institute, Inc., Cary, NC); all other analyses were performed using Stata statistical software (Intercooled Stata 8.0 for Windows, 2003; Stata Corporation, College Station, TX). 
