Received 20 November 2013; returned 19 March 2014; revised 10 April 2014; accepted 19 May 2014

Transcription

1 J Antimicrob Chemother 2014; 69: doi: /jac/dku207 Advance Access publication 11 June 2014 Clinical cure rates in subjects treated with azithromycin for community-acquired respiratory tract infections caused by azithromycin-susceptible or azithromycin-resistant Streptococcus pneumoniae: analysis of Phase 3 clinical trial data George G. Zhanel 1 *, Kevin D. Wolter 2, Cristina Calciu 3, Patricia Hogan 4, Donald E. Low 5, Karl Weiss 6 and James A. Karlowsky 1 1 Department of Medical Microbiology and Infectious Diseases, Faculty of Medicine, University of Manitoba, Winnipeg, Canada; 2 Established Products Clinical Development, Pfizer Inc., New York, NY, USA; 3 Medical Affairs, Pfizer Canada Inc., Montreal, Canada; 4 Specialty Care Business Unit, Pfizer Inc., New York, NY, USA; 5 Department of Microbiology, Mount Sinai Hospital, Toronto, Canada; 6 Maisonneuve-Rosemont Hospital, Faculty of Medicine, University of Montreal, Montreal, Canada *Corresponding author. Microbiology, Health Sciences Centre, MS Sherbrook Street, Winnipeg, Manitoba, R3A 1R9, Canada. Tel: ; Fax: ; ggzhanel@pcs.mb.ca Received 20 November 2013; returned 19 March 2014; revised 10 April 2014; accepted 19 May 2014 Background: Community-acquired respiratory tract infections (CARTI) are commonly caused by Streptococcus pneumoniae (SPN) and empirically treated with azithromycin. This study assessed clinical cure rates in azithromycin-treated subjects with CARTI caused by azithromycin-susceptible (Azi-S) or azithromycin-resistant (Azi-R) SPN. Methods: 1127 subjects with CARTI (402 acute otitis media, 309 community-acquired pneumonia, 255 acute bacterial exacerbations of chronic bronchitis and 161 acute bacterial sinusitis) in 13 Phase 3 clinical trials ( ) had a confirmed pathogen, received azithromycin and were assessed for clinical cure/failure. 34.4% of subjects (388/1127) had a positive culture for SPN; 33.4% (376/1127) had Azi-S or Azi-R SPN. Results: 28.9% (112/388) of subjects with SPN had Azi-R SPN: 35.7% (40/112) were low-level Azi-R SPN (LLAR; MIC 2 8 mg/l), while 64.3% (72/112) were high-level Azi-R SPN (HLAR; MIC 16 mg/l). Among Azi-S and Azi-R SPN CARTI subjects, clinical cure rates were: 86.2% (324/376) overall; 89.4% (236/264) for subjects with Azi-S SPN; 78.6% (88/112) for subjects with Azi-R SPN (P¼0.003, versus Azi-S); 77.5% (31/40) for subjects with LLAR SPN (P, 0.001); and 79.2% (57/72) for subjects with HLAR SPN (P¼0.122). Conclusions: Clinical cure rates in CARTI subjects treated with azithromycin were higher for Azi-S SPN (89.4%) versus Azi-R SPN (78.6%; P¼ 0.003). However, cure rates were not different for subjects infected with LLAR- SPN versus HLAR-SPN. At the observed prevalence of Azi-R SPN of 28.9%, an additional 3.1 clinical failures would be predicted, as a consequence of azithromycin resistance (LLAR and HLAR), per 100 subjects treated empirically with azithromycin. Keywords: resistance, outcome, susceptibility, macrolides, S. pneumoniae Introduction Streptococcus pneumoniae is an important pathogen associated with community-acquired respiratory tract infection (CARTI), including community-acquired pneumonia (CAP), acute bacterial exacerbations of chronic bronchitis (ABECB), acute bacterial sinusitis (ABS) and acute otitis media (AOM). 1,2 Macrolides, especially azithromycin and clarithromycin, are used extensively for the empirical treatment of CARTI due to their broad-spectrum activity against both typical and atypical bacterial pathogens. 2 However, clinical isolates of S. pneumoniae recovered from subjects in North America and worldwide frequently demonstrate resistance to macrolides when tested and interpreted by standardized in vitro methods. 2,3 Resistance to macrolides in pneumococci is generally mediated by one of two mechanisms. 2,4 The first mechanism involves the methylation of a highly conserved region (A2058) of the peptidyltransferase loop in domain V of 23S rrna via an erm(b) methylase. This methylase confers a macrolide (M) lincosamide # The Author Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please journals.permissions@oup.com 2835

2 Zhanel et al. (L) streptogramin B (S B ) resistance (MLS B ) phenotype. 2,4 Typically, the MLS B phenotype is associated with high-level resistance (erythromycin MIC 16 mg/l) to all MLS B antibiotics. 2,4 The second resistance mechanism involves macrolide efflux via mef(a/e) genes, which confer resistance to 14- and 15-membered macrolides only (M phenotype) and not lincosamides and streptogramin B. 2,4 Typically, S. pneumoniae with an M phenotype express low-level resistance (erythromycin MIC 1 8 mg/l) to macrolides. 2,4 The clinical relevance and significance of macrolide-resistant S. pneumoniae has been questioned. 4 9 However, the clinical relevance and clinical outcomes of macrolide-resistant S. pneumoniae have not been well described when subjects with CARTI are treated with azithromycin. The purpose of this study was to compare clinical cure rates in subjects with CARTI caused by azithromycin-susceptible or azithromycin-resistant S. pneumoniae and treated with azithromycin by using data extracted from Phase 3 clinical efficacy trials dealing with CARTI in the Pfizer clinical trials repository. Patients and methods Description of clinical trials The clinical trials selected for review were all pivotal Good Clinical Practice (GCP) Phase 3 randomized comparative or non-comparative clinical trials in Pfizer s repository that enrolled subjects with CARTI who were treated with azithromycin (monotherapy) and had an accessible electronic database; no other suitable CARTI trials in Pfizer s repository were available for inclusion. Eleven of the 13 trials were randomized, comparative, blinded clinical trials, while 2 of 13 were non-comparative (open-label) GCP Phase 3 efficacy trials performed in Japan. Three of the trials were conducted in a single country (USA or Japan); the remainder were international studies predominantly USA/Canada and the European Union, but also including study centres in Central and South America, South Africa, China, India, Israel and Russia. All subjects were enrolled between 1993 and Data and subjects All subjects from the 13 trials were included in the analysis population who (i) had the confirmed pathogen of interest (i.e. S. pneumoniae) with azithromycin MIC (doubling dilution) recorded; (ii) were treated with azithromycin; and (iii) were assessed for clinical cure/failure at test of cure (TOC) according to the individual study protocol; subjects who discontinued from the trial prior to the TOC visit were designated as clinical failures according to the pre-specified requirements of the individual study protocol. Definitions of clinical efficacy Efficacy definitions varied slightly between studies since they involved different infections. However, the general definitions were as follows: clinical cure (signs and symptoms related to the acute infection had resolved or clinical improvement was such that no additional antibiotics were deemed necessary and, for CAP, the chest X-ray performed at the TOC visit had either improved or did not show progression); and failure (one or more of the following occurred: signs and symptoms related to the acute infection had persisted or worsened and additional antibiotics were necessary; new clinical signs and/or symptoms had appeared and additional antibiotics were necessary; and, for CAP, radiological evidence of pneumonia progression during treatment or death due to infection). Antimicrobial susceptibility testing All isolates were tested for antimicrobial susceptibilities using CLSI microdilution methods. Azithromycin-susceptible S. pneumoniae was defined as an isolate with an azithromycin MIC 0.5 mg/l. Low-level azithromycin resistance (LLAR) was defined as an azithromycin MIC of 2 8 mg/l and high-level azithromycin resistance (HLAR) was defined as an azithromycin MIC 16 mg/l. The presence of macrolide resistance genes mef(a/e) or erm(b) was not assessed nor was the presence of an MLS B phenotype assessed based upon susceptibility to other antimicrobials. Data analysis Clinical cure and MIC data were pooled across all included studies in order to categorically compare cure rates between subjects with azithromycinsusceptible or azithromycin-resistant S. pneumoniae, excluding subjects with azithromycin-intermediate S. pneumoniae. The clinical cure rate (%) was expressed as the number of subjects cured/number of subjects treated. Cure rates were also calculated according to susceptibility categories based on the MIC values, using the LLAR and HLAR classifications. Statistical analysis Clinical cure rates were compared across susceptibility categories using the Mantel Haenszel x 2 test, stratifying by study. The proportion of subjects who were azithromycin resistant was compared across indications using a Pearson x 2 test. The predicted clinical cure rate as a function of non-categorized MIC levels was estimated using a logistic regression model with log-transformed MIC as a predictor. Statistical significance was declared when the P value was,0.05. Results A total of 1127 subjects with CARTI met the analysis criteria andincludedbothpaediatric(,16 years, n¼402) and adult ( 16 years, n¼725) subjects. Of the 13 studies, 11 were randomized, comparative and blinded (n¼996 subjects), while 2 of 13 were non-comparative, open-label trials (n¼131 subjects). Five studies enrolled CAP subjects, four studies enrolled ABECB subjects and two studies each enrolled ABS and AOM subjects. Comparator antimicrobials included levofloxacin in five studies, moxifloxacin in two studies, clarithromycin in two studies, amoxicillin in one study and azithromycin in one study (in which two different azithromycin regimens were compared); two studies had no antimicrobial comparator. A total of 388 of 1127 (34.4%) subjects had S. pneumoniae isolated at baseline and are the subject of this report; 12 of 388 (3.1%) isolates demonstrated intermediate azithromycin resistance values. Of the 376 subjects with a documented azithromycin-susceptible or azithromycin-resistant S. pneumoniae respiratory infection who were treated with azithromycin and for whom both clinical cure data and azithromycin MIC data were available, the breakdown included the following: AOM, 47.1% (177/376); CAP, 21.0% (79/376); ABS, 15.2% (57/ 376) and ABECB, 16.7% (63/376) (Table 1). The proportions of subjects with infections caused by azithromycin-resistant S. pneumoniae (mean: 28.9%) were not significantly different (P ¼ 0.191) across the four respiratory infections (range: 22.2% for AOM to 38.5% for ABECB). Results for categorical analyses of clinical cure rate by azithromycin susceptibility (azithromycin susceptible versus azithromycin resistant) are shown for all subjects (n¼376) by resistance 2836

3 Cure rates in subjects treated with azithromycin JAC Table 1. The 376 azithromycin-treated CARTI subjects with azithromycin-susceptible or azithromycin-resistant S. pneumoniae and azithromycin MIC data available for analysis, by indication Indication No. of studies No. of subjects who were culture positive for Azi-S a or Azi-R a S. pneumoniae and received treatment with azithromycin No. of subjects with Azi-R S. pneumoniae Subjects with Azi-R S. pneumoniae (%) b P value c ABECB ABS AOM CAP All subjects b CARTI, community-acquired respiratory tract infection; ABECB, acute bacterial exacerbations of chronic bronchitis; ABS, acute bacterial sinusitis; AOM, acute otitis media; CAP, community-acquired pneumonia. a Azithromycin-resistant (Azi-R) S. pneumoniae defined as an MIC 2 mg/l; azithromycin-susceptible (Azi-S) S. pneumoniae defined as MIC 0.5 mg/l. b Denominators used to calculate frequency of Azi-R isolates include an additional 12 subjects with S. pneumoniae showing intermediately resistant values (MIC.0.5 mg/l and,2 mg/l) not otherwise included in the Azi-S versus Azi-R analyses: CAP, n¼2; ABECB, n¼2; and AOM, n¼8. c Pearson x 2 comparison of proportions of Azi-R S. pneumoniae among subjects with S. pneumoniae, by indication. Table 2. Comparison of clinical cure data for 376 azithromycin-treated subjects with corresponding azithromycin MIC phenotype data a Azithromycin MIC phenotype MIC (mg/l) No. of subjects No. (%) of subjects cured P value b P value c All azithromycin-treated subjects with S. pneumoniae (86.2) Subjects with Azi-S S. pneumoniae (89.4) reference group Subjects with Azi-R S. pneumoniae (78.6) Subjects with LLAR S. pneumoniae (77.5),0.001 reference group Subjects with HLAR S. pneumoniae (79.2) Subjects with very high-level azithromycin-resistant S. pneumoniae (79.3) Azi-S, azithromycin susceptible; Azi-R, azithromycin resistant; LLAR, low-level azithromycin resistant; HLAR, high-level azithromycin resistant. a Analysis included all azithromycin-treated subjects with Azi-S or Azi-R S. pneumoniae in acute bacterial exacerbations of chronic bronchitis, acute bacterial sinusitis, acute otitis media and community-acquired pneumonia studies. b Mantel Haenszel x 2 comparison versus subjects with azithromycin-susceptible S. pneumoniae. c Mantel Haenszel x 2 comparison versus subjects with LLAR S. pneumoniae. phenotype (Table 2) and by type of infection (Table 3). Although the overall clinical cure rate in subjects with CARTI caused by S. pneumoniae and treated with azithromycin was 86.2%, subjects infected with azithromycin-susceptible S. pneumoniae demonstrated a higher clinical cure rate than subjects with azithromycin-resistant isolates (89.4% versus 78.6%, respectively; P¼0.003) (Table 2). However, Table 2 shows that no significant differences in clinical cure rates occurred across a wide range of azithromycin-resistant MIC values, with cure rates for HLAR (azithromycin MIC 16 mg/l; 79.2%) or very high-level azithromycinresistance (azithromycin MIC 64 mg/l; 79.3%) not different compared with the cure rate for LLAR (azithromycin MIC 2 8 mg/l; 77.5%), based on categorical analysis. Within specific respiratory infection types, the clinical cure rate was significantly lower for subjects infected with azithromycinresistant S. pneumoniae isolates, compared with azithromycinsusceptible isolates, for AOM (63.4% versus 83.8%, respectively; P¼0.005) and for ABS (89.5% versus 100%, P¼0.040) (Table 3). For CAP, no difference in the clinical cure rate was observed between subjects infected with azithromycin-susceptible (94.2%) and azithromycin-resistant (92.6%) S. pneumoniae (P ¼ 0.631), while ABECB subjects with azithromycin-resistant S. pneumoniae showed a nominally lower cure rate (80.0% versus 92.1%) that was not statistically different (P¼0.986). While the cure rate for azithromycin-susceptible S. pneumoniae was comparable across the different respiratory infections, a somewhat lower value was observed in the AOM population. The clinical cure rate for all subjects with S. pneumoniae AOM was 79.1% (140/177), with a significant difference in clinical cure rates between azithromycin-susceptible and azithromycin-resistant infections (83.8% versus 63.4%, P¼0.005). The overall clinical impact of azithromycin-resistant S. pneumoniae in a population is a function of the prevalence of azithromycin-resistant isolates in the population and has been expressed here as clinical failures attributable to azithromycinresistant infections/100 empirically treated subjects. In this analysis population, the overall prevalence of azithromycin-resistant S. pneumoniae was 28.9% and the overall observed clinical cure 2837

4 Zhanel et al. Table 3. Clinical outcome data for 376 azithromycin treated subjects with azithromycin-susceptible or -resistant S. pneumoniae, by indication Indication/azithromycin MIC phenotype a Total no. of subjects No. (%) of subjects cured No. (%) of subjects that were clinical failures P value b All subjects azithromycin susceptible (89.4) 28 (10.6) azithromycin resistant (78.6) 24 (21.4) CAP azithromycin susceptible (94.2) 3 (5.8) azithromycin resistant (92.6) 2 (7.4) ABECB azithromycin susceptible (92.1) 3 (7.9) azithromycin resistant (80.0) 5 (20.0) ABS azithromycin susceptible (100) 0 (0) azithromycin resistant (89.5) 2 (10.5) AOM azithromycin susceptible (83.8) 22 (16.2) azithromycin resistant (63.4) 15 (36.6) ABECB, acute bacterial exacerbations of chronic bronchitis; ABS, acute bacterial sinusitis; AOM, acute otitis media; CAP, community-acquired pneumonia. a Azithromycin susceptible, MIC 0.5 mg/l; azithromycin resistant, MIC 2 mg/l; low-level azithromycin resistant, MIC 2 8 mg/l; high-level azithromycin resistant, MIC 16 mg/l. b Mantel Haenszel test for differences in cure rates between susceptibility categories. rate for azithromycin-resistant S. pneumoniae was 78.6%, compared with 89.4% for azithromycin-susceptible isolates. Therefore, the clinical impact of azithromycin resistance can be estimated by considering that if the prevalence of azithromycinresistant S. pneumoniae had been zero, the clinical cure rate in the 28.9% of the population with azithromycin-resistant S. pneumoniae would have been 89.4% rather than 78.6% and there would have been 3.1 additional cures [0.289 ( )] in a population of 100 S. pneumoniae subjects empirically treated with azithromycin. The clinical impact can also be expressed in relation to the positive predictive value (PPV) of antimicrobial susceptibility testing, where PPV¼(number of failures among azithromycin-resistant subjects/number of azithromycin-resistant subjects) 100%. The PPV was only 21% in this study, overall: a CARTI subject with S. pneumoniae having an azithromycin MIC 2 mg/l had a 21% chance of clinical failure and, for CAP, PPV was found to be as low as 7.4% in this study. Discussion The results of our study suggest that there is a relatively weak relationship between the azithromycin MIC and clinical outcome of S. pneumoniae CARTI treated with azithromycin monotherapy: a clinical cure rate that is 10% lower for azithromycin-resistant isolates versus azithromycin-susceptible isolates. As a consequence, the true clinical impact of azithromycin resistance for CARTI appears to be considerably less than the high prevalence of azithromycin resistance in S. pneumoniae might otherwise suggest. The underlying reason for the apparently weak relationship between in vitro susceptibility and clinical cure rate for S. pneumoniae treated with azithromycin cannot be fully explained by these data. The MIC values for the isolates in these studies represent the concentrations of azithromycin necessary to inhibit the growth of S. pneumoniae. While some clinical cures at TOC could be spontaneous, that would not fully explain our overall findings. The clinical cures observed for LLAR S. pneumoniae (azithromycin MIC 2 8 mg/l) and HLAR isolates (azithromycin MIC 16 mg/l) are presumed to be due, at least in part, to azithromycin-resistant pathogens being exposed to a microenvironment characterized by an azithromycin concentration above the MIC, as a result of local lysosomal release of highly concentrated, sequestered azithromycin. 10 Although the North American and global rates of macrolideresistant S. pneumoniae are high (range: 10% to.70%), the clinical impact of these in vitro resistance data has not been fully elucidated. 2,4,5 Clinical and microbiological failures with macrolides have been observed in cases of meningitis, 11 AOM 12 and CAP due to macrolide-resistant pneumococci, 4,13 20 but data are limited. 4,9 Although clinical and microbiological failures with macrolides in the treatment of S. pneumoniae respiratory infections have been reported, the data are primarily from anecdotal case reports and case series. 2,4,9 Wide clinical experience has found that macrolides are highly efficacious for the treatment of CAP 4,5 and our results for azithromycin (monotherapy) used in the treatment of CAP due to S. pneumoniae support that conclusion, irrespective of the MIC for the isolate. So why is there a disparity or lack of predictability between in vitro susceptibility of S. pneumoniae to macrolides and clinical outcomesinpatientswithcarti?doernandbrecher 21 suggested that the clinical predictive value of antimicrobial in vitro susceptibility testing falls into two groups. The first group includes 2838

5 Cure rates in subjects treated with azithromycin JAC single-agent parenteral antimicrobials in the treatment of immunocompetent patients with monomicrobial bacterial infections in circumstances in which the penetration of drug to the site of infection is predictable. In this group there exists the rule, where if the pathogen is susceptible to the antimicrobial a favourable response occurs in 90% 95% of patients, whereas if the organism is resistant a favourable response occurs in 60% of patients. The second group includes oral antimicrobials for the treatment of outpatient infections such as CARTI as well as urinary tract infections, polymicrobial infections and infections treated with multiple antimicrobials, where the results of in vitro susceptibility testing have little to no clinical predictive value. Thus, with oral macrolides for the treatment of CARTI in the outpatient setting, in vitro susceptibility data for S. pneumoniae may have poor or no predictive value, as we found to be the case in our study. The absence of an important association between macrolideresistant S. pneumoniae and clinical failure when treating CARTI with macrolides may be related to the pharmacokinetics/ pharmacodynamics of the drug in cells and tissues. 22,23 In the outpatient setting, infections tend to occur in tissue sites rather than serum and macrolides achieve high concentrations at infection sites or within phagocytes. 10 Concentrations of clarithromycin and azithromycin in lung epithelial lining fluid (ELF) exceed serum concentrations by several times and intracellular concentrations within phagocytes (macrophages or leucocytes) greatly exceed the MIC for S. pneumoniae. 2,10 In addition, macrolide concentrations in ELF (especially azithromycin) have been shown to rise dramatically in response to inflammation. 2,10 Macrolides have also been reported to exert immunomodulatory and anti-inflammatory actions These actions, reported both in vitro and in vivo, are time and dose dependent and could provide beneficial effects in controlling exacerbations of underlying respiratory problems such as bronchiectasis, cystic fibrosis, asthma, panbronchiolitis and cryptogenic organizing pneumonia. 24,27,28 Retrospective analysis of subjects with CAP showed that addition of a macrolide to the treatment regimen significantly reduced mortality compared with no added macrolide. 29,30 A recent double-blind, randomized clinical trial with 200 subjects showed that clarithromycin administered intravenously for 3 days in subjects with ventilator-associated pneumonia (VAP) accelerated the resolution of VAP and weaning from mechanical ventilation in surviving subjects and delayed death in those who died of sepsis. 31 In assessing the effect of clarithromycin on markers of inflammation in these subjects, the authors showed that clarithromycin was able to restore the balance between proinflammatory versus anti-inflammatory mediators in subjects with sepsis, which was accompanied by more efficient antigen presentation and increased apoptosis. 32 Thus, there appear to be two important reasons why MIC values with azithromycin and S. pneumoniae may predict clinical outcome relatively poorly in the treatment of CARTI. First, as an azalide, serum is not the reservoir supplying the tissues with azithromycin. Azithromycin achieves high concentrations at sites of infection as a result of continual delivery of the drug by lysosomes within circulating phagocytic cells, which serve as a source of effective tissue drug concentrations many days after the drug has cleared from the serum. Second, the immunomodulatory and antiinflammatory properties of azithromycin and other macrolides may potentiate the antimicrobial actions of these agents. This study suffers from several limitations or potential biases. Although we selected all pivotal GCP Phase 3 randomized comparative or non-comparative clinical trials in Pfizer s repository that enrolled subjects with CARTI who were treated with azithromycin monotherapy and had an accessible electronic database, there is potential for bias versus those older trials without accessible electronic databases. In addition, the efficacy criteria were not standard between all the studies due to different sites of infection and 2 of the 13 studies were non-comparative, openlabel studies, although these 2 studies had overall results that were generally consistent with the results from the 11 comparative studies. In addition, these clinical trials assessed clinical cure and not microbiological eradication. Assessing the impact of antimicrobial resistance on clinical outcome may not be as sensitive as microbiological outcome. Also, we assessed the impact of azithromycin resistance on clinical outcome in a variety of different respiratory infections, including pneumonia where the immunomodulatory and anti-inflammatory properties of azithromycin may have affected outcome. In these analyses, we only studied the effects of azithromycinresistant S. pneumoniae on clinical outcome, but not other azithromycin-resistant organisms, because the most important macrolide resistance controversy is with S. pneumoniae. The clinical impact of azithromycin resistance in other organisms is currently being investigated. Finally, we did not test the actual mechanisms of macrolide resistance in S. pneumoniae [e.g. mef(a/e) or erm(b)]; rather, we designated strains as azithromycin susceptible or resistant based on the MIC and assessed low-level and high-level resistance only based on MICs. Information on the presence of an MLS B phenotype, based upon susceptibility to other antimicrobials, was not available. In conclusion, while clinical cure rates in subjects with CARTI treated with azithromycin were statistically significantly higher in subjects infected with azithromycin-susceptible S. pneumoniae (89.4%) versus azithromycin-resistant S. pneumoniae (78.6%; P ¼ 0.003), no difference in clinical cure occurred in subjects infected with LLAR S. pneumoniae (MIC 2 8 mg/l) versus HLAR S. pneumoniae (MIC 16 mg/l). The overall clinical impact of azithromycin resistance with S. pneumoniae in CARTI, at the observed prevalence of azithromycin resistance in S. pneumoniae of 28.9%, would be an additional 3.1 clinical failures per 100 subjects empirically treated with azithromycin. Acknowledgements These data were presented in part at the Infectious Diseases Society of America meeting in Boston, MA, October 2011 (Abstract A209). The authors gratefully acknowledge statistical support provided by Pfizer (Dr John Orazem) and the clinical contributions of Dr Alan Kaplan. Funding This study was carried out as part of our routine work. The clinical trials that are the subject of this paper were all funded by Pfizer. Transparency declarations K. D. W, C. C. and P. H. are full-time employees of Pfizer and own Pfizer stock. All other authors have none to declare. 2839

6 Zhanel et al. References 1 Lynch JP, Zhanel GG. Streptococcus pneumoniae: epidemiology risk factors and strategies for prevention. Sem Resp Crit Care Med 2009; 30: Lynch JP, Zhanel GG. Streptococcus pneumoniae: antimicrobial resistance. Curr Opin Pulm Med 2010; 16: Bishai W. A testament to sustained macrolide efficacy. Clin Infect Dis 2003; 36: Rzeszutek M, Wierzbowski A, Conly J et al. A review of clinical failures involving macrolide-resistant Streptococcus pneumoniae. Int J Antimicrob Agents 2004; 24: Lynch JD, Martinez FJ. Clinical relevance of macrolide-resistant S. pneumoniae for community acquired pneumonia (CAP). Clin Infect Dis 2002; 34 Suppl 1: S Bishai WR. Clinical significance of pneumococcal resistance and factors influencing outcomes. Treat Respir Med 2005; 4 Suppl 1: Nuermberger E, Bishai WR. The clinical significance of macrolideresistant Streptococcus pneumoniae: it s all relative. Clin Infect Dis 2004; 38: Nuermberger EL, Bishai WR. Antibiotic resistance in Streptococcus pneumoniae: what does the future hold? Clin Infect Dis 2004; 38 Suppl 4: Bishai W. The in vivo in vitro paradox in pneumococcal respiratory tract infections. J Antimicrob Chemother 2002; 49: Zhanel GG, Dueck M, Hoban DJ et al. Macrolides and ketolides: a review focusing on respiratory infections. Drugs 2001; 61: Reid R, Bradley JS, Hindler J. Pneumococcal meningitis during therapy of otitis media with clarithromycin. Pediatr Infect Dis 1995; 14: Dagan R, Leibovitz E, Leiberman A et al. Clinical significance of antibiotic resistance in acute otitis media and implication of antibiotic treatment on carriage and spread of resistant organisms. Pediatr Infect Dis 2000; 19 Suppl 5: Waterer GW, Wunderink RG, Jones CB. Fatal pneumococcal pneumonia attributed to macrolide resistance and azithromycin monotherapy. Chest 2000; 118: Kelley MA, Weber DJ, Gilligan P et al. Breakthrough pneumococcal bacteremia in subjects being treated with azithromycin and clarithromycin. Clin Infect Dis 2000; 31: Fogarty C, Goldschmidt R, Bush K. Bacteremic pneumonia due to multidrug-resistant pneumococci in 3 subjects treated unsuccessfully with azithromycin and successfully with levofloxacin. Clin Infect Dis 2000; 31: Lonks JR. What is the clinical impact of macrolide resistance? Curr Infect Dis Rep 2004; 6: Moreno S, Garcia-Leoni ME, Cercenado E et al. Infections caused by erythromycin-resistant Streptococcus pneumoniae: incidence, risk factors, and response to therapy in a prospective study. Clin Infect Dis 1995; 20: Lonks JR, Garau J, Gomez L et al. Failure of macrolide antibiotic treatment in subjects with bacteremia due to erythromycin-resistant Streptococcus pneumoniae. Clin Infect Dis 2002; 35: Van Kerkhoven D, Peetermans WE, Verbist L et al. Breakthrough pneumococcal bacteraemia in subjects treated with clarithromycin or oral b-lactams. J Antimicrob Chemother 2003; 51: Musher DM, Dowell ME, Shortridge VD et al. Emergence of macrolide resistance during treatment of pneumococcal pneumonia. N Engl J Med 2002; 346: Doern GV, Brecher SM. The clinical predictive value (or lack thereof) of the results of in vitro antimicrobial susceptibility tests. J Clin Microbiol 2011; 49: Liu P, Allaudeen H, Chandra R et al. Comparative pharmacokinetics of azithromycin in serum and white blood cells of healthy subjects receiving a single-dose extended-release regimen versus a 3-day immediate-release regimen. Antimicrob Agents Chemother 2007; 51: Gerbaux C, Van Bambeke F, Montenez J-P et al. Hyperactivity of cathepsin-b and other lysosomal enzymes in fibroblasts exposed to azithromycin, a dicationic macrolide antibiotic with exceptional tissue accumulation. FEBS Lett 1996; 394: Zarogoulidis P, Papanas N, Kioumis I et al. Macrolides: from in vitro antiinflammatory and immunomodulatory properties to clinical practice in respiratory diseases. Eur J Clin Pharmacol 2012; 68: Giamarellos-Bourboulis EJ, Raftogiannis M. The immune response to severe bacterial infections: consequences for therapy. Expert Rev Anti Infect Ther 2012; 10: Kovaleva A, Remmelts HH, Rijkers GT et al. Immunomodulatory effects of macrolides during community-acquired pneumonia: a literature review. J Antimicrob Chemother 2012; 67: Mammen MJ, Sethi S. Macrolide therapy for the prevention of acute exacerbations in chronic obstructive pulmonary disease. Pol Arch Med Wewn 2012; 122: Good JT Jr, Rollins DR, Martin RJ. Macrolides in the treatment of asthma. Curr Opin Pulm Med 2012; 18: Asadi L, Eurich DT, Gamble JM et al. Guideline adherence and macrolides reduced mortality in outpatient subjects with pneumonia. Respir Med 2012; 106: AsadiL, SliglWI, EurichDTet al. Macrolide-based regimens and mortality in hospitalized subjects with community-acquired pneumonia: a systematic review and meta-analysis. Clin Infect Dis 2012; 55: Giamarellos-Bourboulis EJ, Pechere JC, Routsi C et al. Effect of clarithromycininsubjectswithsepsisandventilatory-associatedpneumonia. Clin Infect Dis 2008; 46: Spyridaki A, Raftogiannis M, Antonopoulou A et al. Effect of clarithromycin in inflammatory markers of subjects with ventilatory-associated pneumonia and sepsis caused by Gram-negative bacteria: results from a randomized clinical study. Antimicrob Agents Chemother 2012; 56: