Antimicrobial resistance determinants and susceptibility profiles of pneumococcal isolates recovered in Trinidad and Tobago

Highlights • AWGS-based approach can accurately and reliably predict antimicrobial phenotypes.• Observed rates of non-susceptibility to SXT and erythromycin were lower than reported for other countries in the region.• In contrast,the proportion of β-lactam non-susceptibility was higher compared with other countries in the region.• Multidrug resistance remains low but appears to be expanding clonally following introduction of pneumococcal vaccines.• This clonal expansion is driven by the 19F-CC156 and 19A/F-CC236 lineages.


Introduction
Infections caused by Streptococcus pneumoniae include diseases such as meningitis, bacteraemia and pneumonia as well as less severe conditions such as sinusitis and otitis media. The World Health Organization (WHO) estimated that pneumococcal infections caused 476 000 (5%) deaths globally among human immunodeficiency virus (HIV)-negative children under 5 years of age during 2008 [1]. In Latin America and the Caribbean, pneumococcal infections were estimated to account for 12 000-18 000 deaths, 327 000 cases of pneumonia, 4000 cases of meningitis and 1229 cases of sepsis each year in children under five years old [2].
The first pneumococcal conjugate vaccine (PCV) covered seven serotypes (PCV7: serotypes 14, 6B, 19F, 23F, 4, 9V and 18C) and was licensed in 2000, followed by PCV10 (PCV7 serotypes plus 1, 5 and 7F) in 2009, and PCV13 (PCV10 serotypes plus 3, 6A and 19A) in 2010. In the Caribbean, as well as in other regions, vaccine uptake has been variable. PCV7 was introduced into the national immunisation programme (NIP) in Trinidad and Tobago in March 2010 for infants at risk of pneumococcal disease, mainly those with immune deficiencies and other chronic diseases. Prior to 2010, $ The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the US Centers for Disease Control and Prevention (CDC). PCV7 was only available in the private sector [3]. PCV10 was introduced into the NIP in 2011 for all children aged <2 years [3][4][5] and was replaced by PCV13 in August 2015. As of December 2014, vaccine coverage with PCV10 was reported to be 95% [6].
From the beginning of the antibiotic era to the mid-1970s, S. pneumoniae remained uniformly susceptible to all classes of antibiotics that were active against it, with the exception of tetracycline. In the ensuing decades, resistance of pneumococci to a variety of antimicrobials has evolved into a worldwide health problem [7]. A 2004 report by the SENTRY surveillance program showed that penicillin non-susceptibility rates were as high as 25% among pneumococcal isolates, with penicillin-non-susceptible isolates presenting higher rates of multidrug resistance [8].
In the Caribbean region, infections due to penicillin-resistant pneumococci have been reported in hospitals and community settings, but limited data are available to estimate patterns of drug resistance [9]. The aim of this study was to determine the antimicrobial susceptibility profiles of 98 pneumococcal isolates collected in Trinidad and Tobago from invasive and non-invasive sites and their associated genetic determinants.

Materials and methods
A total of 98 pneumococcal isolates recovered at several regional hospitals in Trinidad and Tobago were included in the study, including 83 invasive and 15 non-invasive strains. All pneumococcal isolates (n = 73) obtained from routine clinical specimens at the five major public hospitals in Trinidad and Tobago during the period 2011-2013 were included in this study. A number of historical pneumococcal isolates (n = 25) from clinical specimens that were collected between 1997-2010 were also included in the analysis; most of the strains were collected from three regional hospitals prior to the start of the SIREVA (Sistema regional de vacunas) project in Trinidad and Tobago. Serotyping and multilocus sequence typing (MLST) results from these 98 isolates have been published previously [10]. Minimum inhibitory concentrations (MICs) were determined using the broth microdilution method as specified by Clinical and Laboratory Standards Institute (CLSI) guidelines [11]. S. pneumoniae strains were cultured on BD TM Trypticase TM Soy Agar II with 5% sheep blood (Beckton Dickinson, Heidelberg, Germany) and incubated overnight at 37 C in 5% CO 2 . Genomic DNA was then extracted manually using a modified QIAamp 1 DNA Mini Kit (QIAGEN, Inc., Valencia, CA) protocol. Whole-genome sequencing (WGS) was performed at the Sanger Institute using an Illumina HiSeq 2500 system (Illumina Inc., San Diego, CA) as part of the Global Pneumococcal Sequencing Project (http://www.pneumogen.net) and the resulting data were submitted to the European Nucleotide Archive (accession nos. are given in Supplementary  Table S1). Sequences were analysed using the Streptococcus laboratory pneumococcal typing pipeline of the US Centers for Disease Control and Prevention (CDC) [12,13] to identify core genomic alterations and accessory elements that confer antimicrobial resistance phenotypes as well as pilus genes (https://github.com/BenJamesMetcalf/Spn_Scripts_Reference).
Contingency tables and a x 2 test (or Fisher's exact test) were used to determine the significance of associations (at a = 0.05).

Antimicrobial resistance
The capability of a WGS-based approach to accurately and reliably predict antimicrobial phenotypes has been previously shown to be an adequate substitute for broth dilution susceptibility testing [13,14]. By a WGS-based assessment of resistance, 34 (34.7%) of the 98 isolates in this study were predicted to be non-susceptible to trimethoprim/sulfamethoxazole (SXT), 18 (18.4%) resistant to erythromycin, 17 (17.3%) non-susceptible to b-lactams, 9 (9.2%) resistant to tetracycline, 2 (2.0%) resistant to chloramphenicol and 1 (1.0%) resistant to rifampicin. These results were in agreement with the MICs determined by broth microdilution (Table 1). Overall, 41 isolates (41.8%) were predicted to be resistant to at least one antimicrobial class, including 13 (13.3%) resistant to at least three classes [i.e. multidrug-resistant (MDR)]. Before PCV7/10 introduction, 28.0% (7/25) of the isolates were resistant to at least one antimicrobial class; this proportion increased to 46.6% (34/73) in the post-PCV period (p = 0.06). The proportion of MDR isolates increased from 4.0% (1/25) to 16.4% (12/73) after PCV7/10 introduction (p = 0.06). A total of 37 PBP allele combinations (PBP types) were identified among these isolates, 11 of them novel; 8 (21.6%) of these combinations (3 of them novel) were associated with non-susceptibility to penicillin (MIC 0.12 mg/L). Identifying new allele combinations was expected as the database used for analyses mostly contains isolates from the USA. Of the 17 isolates that were predicted as non-susceptible to b-lactams, all were predicted as penicillin non-susceptible (PNS), 9 were predicted as non-susceptible to ceftriaxone, cefuroxime, cefotaxime and meropenem (in addition to penicillin), and 3 were predicted as non-susceptible to all six of the b-lactams tested.
Of the 34 isolates predicted as non-susceptible against co-trimoxazole, all contained one to two codon insertions within the folP gene (intermediate phenotype, MIC 1-2 mg/L), whilst 19 (55.9%) also contained the I100L substitution in folA (resistant phenotype, MIC 4 mg/L). Of the 18 isolates predicted as resistant to erythromycin, 11 (61.1%) were positive for mef alone, 2 (11.1%) for ermB alone and 5 (27.8%) for ermB + mef; the latter 7 isolates containing ermB were also predicted as resistant to clindamycin. In addition, nine isolates were positive for tetM and two isolates for the cat gene. One isolate contained a change in the rpoB gene (H499Y) and was resistant to rifampicin by broth microdilution (MIC > 2 mg/L).
The observed rate of non-susceptibility against co-trimoxazole (34.7%) was lower than what has been reported for other countries in the region [15] such as Venezuela (100%) and the Dominican Republic (65%); the rate of resistance to erythromycin (18.4%) was similarly lower (compared with 45% in Venezuela and 20% in Dominican Republic). In contrast, the proportion of b-lactam nonsusceptibility (17.3%) was higher among these isolates than among isolates from the Dominican Republic (9.6%), but was similar to that observed among isolates recovered in Venezuela (18.2%).
In conclusion, this study offers a snapshot of the antimicrobial resistance profiles and genetic resistance determinants among 98 pneumococcal isolates recovered in Trinidad and Tobago, adding to the limited body of data available for the Caribbean region. The observed rates of resistance were similar to those reported for neighbouring Caribbean countries. Multidrug resistance remains low but appears to be expanding clonally following PCV7/10 introduction, driven by the 19F-CC156 and 19A/F-CC236 lineages. Thus, the introduction of PCV13 will likely have a marked impact on pneumococcal multidrug resistance in Trinidad and Tobago.

Competing interest
None declared.

Funding
Isolates were characterised as part of the Global Pneumococcal Strain Bank established with funding from PATH and currently housed at the US Centers for Disease Control and Prevention (CDC) (https://www.cdc.gov/streplab/global-pneumo-strain-bank.html). Whole-genome sequencing was performed as part of the Global  Pneumococcal Sequencing Project, funded by the Bill and Melinda Gates Foundation [grant no. OPP1034556]. The funding sources had no involvement in the study design, the collection, analysis or interpretation of data, the writing of the report, or the decision to submit the article for publication.

Ethical approval
Not required.