Regional and global antimicrobial susceptibility among isolates of Streptococcus pneumoniae and Haemophilus influenzae collected as part of the Tigecycline Evaluation and Surveillance Trial (T.E.S.T.) from 2009 to 2012 and comparison with previous years of T.E.S.T. (2004-2008)

Background We report here on 14438 Streptococcus pneumoniae and 14770 Haemophilus influenzae isolates collected from 560 centres globally between 2004 and 2012 as a part of the Tigecycline Evaluation and Surveillance Trial (T.E.S.T.). Methods MIC testing was performed using broth microdilution methods as described by the Clinical and Laboratory Standards Institute (CLSI) using CLSI-approved breakpoints; US Food and Drug Administration breakpoints were used for tigecycline as CLSI breakpoints are not available. Results At least 99% of S. pneumoniae isolates globally were susceptible to levofloxacin, linezolid, tigecycline or vancomycin. Penicillin resistance was observed among 14.8% of S. pneumoniae and was highest in Asia/Pacific Rim (30.1%) and Africa (27.6%); 23.4% of S. pneumoniae isolates were penicillin-intermediate, which were most common in Africa (37.6%). Minocycline susceptibility among S. pneumoniae decreased by 20% between 2004-2008 and 2009-2012. High (>98.5%) susceptibility was reported among H. influenzae to all antimicrobial agents on the T.E.S.T. panel excluding ampicillin, to which only 78.3% were susceptible. β-lactamase production was observed among 20.2% of H. influenzae isolates; 1.5% of isolates were β-lactamase negative, ampicillin-resistant. Conclusions S. pneumoniae remained highly susceptible to levofloxacin, linezolid, tigecycline and vancomycin while H. influenzae was susceptible to most antimicrobial agents in the testing panel (excluding ampicillin). Electronic supplementary material The online version of this article (doi:10.1186/s12941-014-0052-2) contains supplementary material, which is available to authorized users.


Background
Streptococcus pneumoniae and Haemophilus influenzae are among the most common causes of pneumonia worldwide [1,2]. According to the most recent guidelines for the treatment of community-acquired pneumonia (CAP) published by the Infectious Diseases Society of America and American Thoracic Society [3], the recommended empiric therapy for inpatient CAP includes a β-lactam (cefotaxime, ceftriaxone or ampicillin-sulbactam) plus either a respiratory fluoroquinolone or azithromycin for intensive care unit (ICU) patients or a respiratory fluoroquinolone or β-lactam plus a macrolide for non-ICU patients. European guidelines are also available [4], while several European countries have established their own policies regarding the treatment of CAP including the UK [5], Sweden [6] and the Netherlands [7]. These guidelines often differ in their treatment recommendations, even when bacteriology is similar between countries [8].
The Tigecycline Evaluation and Surveillance Trial (T.E.S.T.) is a global surveillance study which has been ongoing since 2004. This study was designed to monitor the longitudinal activity of a broad panel of comparator agents, including the glycylcycline antimicrobial tigecycline, against a collection of clinically important pathogens. Tigecycline possesses potent in vitro activity against those pathogens most often responsible for community-acquired pneumonia [9]. In 2009, the US Food and Drug Administration approved tigecycline for the treatment of community-acquired bacterial pneumonia (CABP) caused by Streptococcus pneumoniae (penicillin-susceptible isolates), including cases with concurrent bacteremia, Haemophilus influenzae (β-lactamase negative isolates), and Legionella pneumophila in the USA [10], based in part on results from a pair of Phase III clinical trials comparing the efficacy of tigecycline and levofloxacin in hospitalized patients with CABP [11]. In the current study, we examine susceptibility trends among isolates of S. pneumoniae and H. influenzae collected from Africa, the Asia/Pacific Rim, Europe, Latin America, the Middle East and North America between 2004 and 2012.
This report updates S. pneumoniae and H. influenzae data previously reported from T.E.S.T. between 2004 and 2008 [12]. In that report, a total of 6785 S. pneumoniae and 6642 H. influenzae were analyzed. Penicillin resistance among S. pneumoniae ranged from 9.3% in Europe to 25.1% in the Asia-Pacific Rim and β-lactamase producing H. influenzae ranged from 8.7% in South Africa to 26.8% in the Asia-Pacific Rim. The MIC 90 s for tigecycline against S. pneumoniae and H. influenzae were ≤0.12 mg/L and ≤2 mg/L, respectively. Data from that 2004-2008 report are included in the dataset used in this report. This report aims to provide an update to Darabi et al. [12], comparing 2004-2008 data with data collected between 2009 and 2012.

Isolate collection
Between 2004 and 2012, a total of 560 centres globally contributed isolates to the T.E.S.T. study in at least a single study year (Table 1). All centres participating in T.E.S.T. were required to contribute a minimum of 65 Gram-positive and 135 Gram-negative isolates per study year; these were to include at least 15 isolates of both S. pneumoniae and H. influenzae annually. North America and Europe were the main contributors, with 221 and 196 centres participating in the two regions, respectively.
All body sites were considered acceptable sources for isolate collection; a maximum of 25% of isolates in any year could be urinary in origin. Banked or stored isolates were not accepted, nor were duplicate isolates from a single patient. Isolates were collected from 13 different body sites (Table 2); a small number of isolates were collected from medical instruments. Isolates were included into the study if they were considered a clinically relevant causative organism and the probable causative agent of the infection. Patient age, medical history or gender were not considered relevant. Isolate identification was carried out by contributing centres using routine local methodologies.
The central repository for all isolates was International Health Management Associates (IHMA, Schaumburg, IL, USA), who were responsible for organism transport and collection, confirmation of isolate identification and the management of a database including all isolate data. Quality control testing was carried out daily using S. pneumoniae ATCC 49619, H. influenzae ATCC 49247 and H. influenzae ATCC 49766. Approximately 10-15% of isolates were randomly tested annually by IHMA to verify isolate identity and MICs; MIC data were used only if daily QC results were within acceptable ranges as published by the Clinical and Laboratory Standards Institute (CLSI) [13]. Antimicrobial susceptibility was determined using CLSI-approved breakpoints [13] with the exception of tigecycline, for which US Food and Drug Administration (FDA)-approved breakpoints were used as CLSI breakpoints are not available for tigecycline [15]. Penicillin oral breakpoints (susceptible ≤0.06 mg/L, resistant ≥2 mg/L) were used for S. pneumoniae in this study; penicillinintermediate S. pneumoniae (PISP) was defined as a penicillin MIC between 0.12 and 1 mg/L.
The MIC results for 2004-2008 presented here do not exactly match the results presented in Darabi et al. [12] due to retesting of selected isolates from 2004-2008 through the intervening study period or the deletion of untestable isolates from the T.E.S.T. database (most often due to loss of culture viability).
Antimicrobial resistance determination β-lactamase production among H. influenzae was determined at collecting centres using preferred local methodologies. The central lab, IHMA, retested all H. influenzae if the ampicillin results did not agree with the β-lactamase determination. The central laboratory (IHMA) uses the cefinase test [15]. Only 15 β-lactamase-negative, ampicillin-resistant (BLNAR) H. influenzae isolates were collected globally during this study; therefore, we have replicated the approach of Darabi et al. [12] here and have included β-lactamase-negative, ampicillinintermediate (BLNAI) isolates as well. This resulted in a total of 228 BLNAR isolates globally.

Statistical analysis
A comparison of rates of susceptibility between the two time periods 2004-2008 and 2009-2012 was carried out using the Fisher's Exact Test. The upper and lower 95% confidence intervals (CI) presented in this manuscript relate to percentage susceptibility. The 95% CI for the percentage susceptible presented in this manuscript were calculated using the SAS exact method for binomial confidence intervals.

H. influenzae
BLNAR isolates were collected from all regions globally and were least common in Middle East (1.0%) and North America (1.1%); prevalence was highest in Asia/ Pacific Rim (3.5%). The global incidence of BLNAR remained consistent during this study, at 1.5% in 2004-2008 and 1.6% in 2009-2012 (Table 3). Antimicrobial susceptibility also did not change over the study period (Additional file 1: Table S2).

Discussion
This report shows that S. pneumoniae global susceptibility has remained relatively stable to most antimicrobial agents globally during the period 2004 to 2012. This reflects the findings of the European Antimicrobial Resistance Surveillance Network (EARS-Net), who showed that non-susceptibility among invasive S. pneumoniae in Europe has remained stable between 2004 and 2011, although large variations existed between countries [1]. Results from the SENTRY study in the USA between 1998 and 2011 showed increased resistance among S. pneumoniae to amoxicillin-clavulanate, penicillin (at MIC ≥4 mg/L) and ceftriaxone (18.9%, 14.8% and 11.7%, respectively), although levofloxacin, linezolid, tigecycline and vancomycin retained 100% activity [16]. Differences between this study and that of Jones et al. [16] may be attributed to the different time period of isolate collection (SENTRY commenced in 1998 compared with the start of TEST in 2004) as well as the source of isolate collection. Also, geography may influence the findings as the current study is a global collection of isolates compared with the USA analysis of Jones et al. [16]. The change in the CLSI breakpoints for S. pneumoniae and penicillin in 2008 also mean that caution must be exercised when comparing susceptibility data between studies [17]. The Jones et al. [16] study used an MIC of ≥4 mg/L for penicillin resistance whereas in this T.E.S.T. study we have used the CLSI penicillin oral breakpoints where resistance is defined as an MIC of ≥2 mg/L. Susceptibility to ceftriaxone and minocycline among S. pneumoniae isolates decreased in all the regions; although in the case of ceftriaxone rates of susceptibility remained high. Reasons for these decreases are unknown and warrant further investigation. One factor that may influence susceptibility among S. pneumoniae is the use of the seven-and 13-valent pneumococcal conjugate vaccines (PCV7 and PCV13) which have been shown to modify the epidemiology of pneumococcal disease [18]. The S. pneumoniae collected as part of T.E.S.T. which have elevated ceftriaxone and minocycline MICs warrant molecular analysis to investigate the influence of these vaccines on the resistance profile of S. pneumoniae globally.
The high levels of susceptibility to tigecycline among S. pneumoniae and H. influenzae reported in the current study mirror recent results presented elsewhere. The Assessing Worldwide Antimicrobial Resistance Evaluation Program (AWARE) reported 100% tigecycline susceptibility among Latin American S. pneumoniae isolates from respiratory tract or complicated skin and soft tissue infections [19] The tigecycline susceptibility results presented here for S. pneumoniae are different from those presented by Darabi et al. [12] in their discussion of S. pneumoniae collected globally between 2004 and 2008 as a part of the T.E.S.T. study. In the earlier report, S. pneumoniae susceptibility to tigecycline ranged from 87.5% in Latin America to 97.1% in the Middle East; in the current report, tigecycline susceptibility was ≥99.8% in all global regions. These changes are likely due to the MIC retesting of several hundred reportedly tigecycline nonsusceptible pneumococci since the publication of Darabi et al. [12]. Re-testing occurred because of the determination that lysed sheep blood in the testing media results in artificially high tigecycline MICs. The re-testing program is carried out by the central laboratory, IHMA, who use un-lysed sheep blood to retest all S. pneumoniae. This program began with the retesting of the most recent isolates and then testing backwards to the first isolates received and therefore, some early S. pneumoniae MIC results may have been published before all retests were completed. To date, all S. pneumoniae with a tigecycline MIC above 0.06 mg/L have retested at or below 0.06 mg/L (data not shown). The difference in results may also be due to the deletion of unevaluable isolates from the T.E.S.T. database. During the production of the dataset used in this manuscript, 236 quarantined S. pneumoniae isolates could not be re-evaluated for susceptibility (i.e., isolates which died on transport from collecting centre to the central laboratory [IHMA], were never shipped to IHMA, or could not be resuscitated for retesting) and were thus permanently removed from the T.E.S.T. database (S. Bouchillon, pers com).
As discussed above, in the current study, oral breakpoints (susceptible ≤0.06 mg/L, resistant ≥2 mg/L) have been used for penicillin against S. pneumoniae and not the parenteral breakpoints (meningitis: susceptible ≤0.06 mg/L, resistant ≥0.12 mg/L; non-meningitis: susceptible ≤2 mg/L, resistant ≥8 mg/L). Use of these oral breakpoints in the current study suggests that penicillin resistance is high globally (PISP, 23.4%; PRSP, 14.8%), thus oral penicillin treatment of pneumonia caused by S. pneumoniae would appear to be ill-advised. However, the occurrence of S. pneumoniae isolates with MICs ≥4 mg/L is rare worldwide [22], and the use of high-dose parenteral penicillin is generally effective in the treatment of pneumonia caused by drug-resistant S. pneumoniae [23].
H. influenzae isolates were highly susceptible to most agents on the T.E.S.T. panel, with global susceptibility ranging from 98.8% (tigecycline) to 99.9% (ceftriaxone, levofloxacin, imipenem and meropenem) during the 2004-2012 interval; the sole exception was ampicillin, to which only 78.3% of isolates were susceptible globally. Regionally, susceptibility in Europe (83.6%) was higher than in North America (73.4%). These regional results compare well with Jones et al., who reported 85.2% and 71.7% ampicillin susceptibility among H. influenzae isolates from Europe and the USA, respectively [24]. Global BLNAR levels were low (1.5%) in the current study. Japan is known to have a relatively high prevalence of BLNAR H. influenzae [25]; however, no Japanese sites contributed isolates to this study. The USA and Europe, where the majority of T.E.S.T. isolates were collected have lower rates but rates have been reported to be increasing in Europe [26,27]. Studies contemporary with this T.E.S.T. study are scarce making direct comparisons of percentages difficult; however, continued monitoring of these isolates is important, particularly given reports of increasing prevalence.
In the first three years of the T.E.S.T. study, contributing centres in North America were two-to three-fold more numerous than centres from Europe. Since 2008, however, European centres have outnumbered those in North America. This is reflected in the numbers of isolates contributed from these two regions: between 2004 and 2008, 48.3% of S. pneumoniae and 44.9% of H. influenzae isolates were contributed by centres in North America compared to 31.4% and 34.5%, respectively, by centres in Europe. When the total study period (2004-2012) is considered, S. pneumoniae is contributed equally from both regions (40.4% from North America compared to 40.0% from Europe); H. influenzae isolates from Europe (42.8%), however, outnumbered those from North America (37.9%). This shift in prevalence from North America to Europe will undoubtedly have influenced overall susceptibility rates observed over the course of this study, as susceptibility was lower in North America than in Europe for most antimicrobial agents between 2004 and 2012. The number of isolates submitted from the other four regions was similar between the two time periods, contributing approximately 19% of isolates. The change in the number of North American and European centres over time reinforces the importance of presentation of regional data from global surveillance studies, where possible. This inconsistency in participating centres over time is an inherent weakness of longitudinal surveillance studies, especially international or global surveillance studies. Thus, as is always the case with data derived from multicentre longitudinal surveillance, the trends in antimicrobial susceptibility reported here must be regarded with some caution.

Conclusions
S. pneumoniae remained highly susceptible to levofloxacin, linezolid, tigecycline and vancomycin while H. influenzae was susceptible to most antimicrobial agents in the T.E.S.T. panel (excluding ampicillin). Tigecycline was approved for the treatment of communityacquired bacterial pneumonia in the USA by the FDA in 2009. The results presented in this report show that the in vitro susceptibility of S. pneumoniae and H. influenzae, two important pathogens which are commonly associated with community-associated pneumonia, has remained high globally in the period since this approval was granted. Tigecycline may thus continue to be an important tool in the armamentarium of physicians for the treatment of CABP.

Additional file
Additional file 1: