Pathogen surveillance and antimicrobial resistance surveillance informs effective antimicrobial prescribing, in addition to controlling the spread of resistant pathogens within the local environment. It is well known that dedicated surveillance systems form an integral strategy to combat antibiotic resistance; however, surveillance of antibiotic-resistant bacteria is severely lacking in most African countries. The largest antimicrobial resistance surveillance data set from sub-Saharan Africa was published in 2017 and reflected antimicrobial resistance trends from isolates implicated in bloodstream infections from adults and children in Malawi over a 19-year period were described.⁹ There have been no long-term surveillance data reports describing ESKAPE pathogens and their antimicrobial resistance patterns from KwaZulu-Natal, South Africa, except for national sentinel site surveillance describing antimicrobial resistance among K. pneumoniae isolates from bloodstream infections from 2009–2012. Hence, this study aimed at providing a description of ESKAPE pathogens and their antibiotic resistance trends over a five-year period in the province of KwaZulu-Natal, South Africa.

The ESKAPE group of bacteria are among the critical and high priority pathogens on the global PPL. This study provides the largest data set of the ESKAPE pathogens gathered from adults and children who presented to the largest public service hospitals in KwaZulu-Natal, South Africa, over a five-year period. The surveillance data provided in this manuscript provide a good baseline surveillance intervention to combat antimicrobial resistance.

The WHO global PPL identifies priority pathogens for the research and development of new antibiotics. Knowledge of these pathogens and their resistance patterns in local contexts is pivotal in our fight against antimicrobial resistance. Despite a decreasing trend in methicillin-resistant S. aureus (MRSA), there is emergence and rapid expansion of extended spectrum beta-lactamase (ESBL)-producing K. pneumoniae coupled with fluoroquinolone resistance observed in this study. Additionally, increasing resistance to third-generation cephalosporins and amoxicillin-clavulanate is a growing concern, as there is now a greater reliance on the carbapenems to treat infections caused by K. pneumoniae. Increasing resistance to carbapenems observed in K. pneumoniae is slowly becoming a reality in KwaZulu-Natal and may spell the beginning of the end for carbapenems against Enterobacteriaceae such as K. pneumoniae. Continued surveillance will provide valuable information on antimicrobial resistance patterns for ESKAPE pathogens, which will assist in informing empiric antimicrobial therapy.

This study was approved by the Biomedical Research Ethics Committee of the University of KwaZulu-Natal (study approval number: BE085/12).

Data collected from August 2011 to December 2015 at nine participating public sector hospitals from two districts across the province of KwaZulu-Natal were utilised in this study. The nine participating hospitals are the largest in the province and represent all levels of healthcare as described in Tables 1 and 2.¹⁰

Each participating centre extracted data of all isolates categorised as ESKAPE pathogens. All body sites were considered acceptable sources. Samples submitted for microbiological analysis included blood, urine, catheters (central venous catheter + hemodialysis) and respiratory specimens, in addition to samples from other body sites. Not every patient who presented to a healthcare facility with an infection submitted a sample for microbiological analysis. Additionally, the data presented included organisms from all specimen types; hence, colonisation was not distinguished from infection.

All participating laboratories in this study subscribed to the National Health Laboratories Services (NHLS) Proficiency Testing Scheme. Evaluations are carried out quarterly and samples for the scheme are prepared with assistance from the National Institute for Communicable Diseases – Centre for Opportunistic, Tropical and Hospital Infections, a division of the NHLS. Thus, all laboratories participating in this study were proficient in testing bacterial samples. All participating laboratories were responsible for sample processing, including isolate identification and susceptibility testing. Pathogen identification was determined using the Vitek 2 (bioMerieux, Marcy l’Étoile, France) platform. Antimicrobial susceptibility testing and minimum inhibitory concentration determinations were performed using the Vitek 2 (bioMerieux, Marcy l’Étoile, France) platform. The results were interpreted according to the criteria of the Clinical Laboratory Standards Institute.¹¹ The antimicrobial test panel included: penicillin, ampicillin, amoxicillin-clavulanate, ceftriaxone, cefepime, cefuroxime, cefoxitin, ceftazadime, imipenem, meropenem, ertapenem, piperacillin-tazobactam, amikacin, gentamicin, erythromycin, clindamycin, linezolid, teicoplanin, vancomycin, fusidic acid, mupirocin, tetracycline, oxacillin, rifampicin, nitrofurantoin, colistin, trimethoprim/sulfamethoxazole, ciprofloxacin and tigecycline. Methicillin resistance in S aureus was inferred by oxacillin resistance.¹¹ The Vitek 2 AST-N255 card was used to perform antimicrobial susceptibility testing on Gram-negative, whereas the AST-P603 card was used for Gram-positive organisms. Presumptive ESBL production in K. pneumoniae was determined using the minimum inhibitory concentration for ceftriaxone and ceftazidime as per the Clinical Laboratory Standards Institute guidelines.¹¹ Isolates were characterised as susceptible or resistant using Clinical Laboratory Standards Institute-approved breakpoints.¹¹

All information regarding specimen collection, sampling and patient laboratory results is deposited on the NHLS laboratory database. The use of this database and access to patient information is restricted to laboratory staff working within the NHLS. Thus, data collection was done at the NHLS and entailed extraction of isolate information from this computerised laboratory database. The data received were de-duplicated, anonymised and patient confidentiality was maintained at all times. Isolate information, specimen type and results of antimicrobial susceptibility testing, including minimum inhibitory concentration data, were also extracted. Once the data were extracted, trends in the total number of ESKAPE pathogens and their antimicrobial resistance patterns were determined. These trends were compared over a five-year period.

All data processing and analyses were performed using Stata 13.0 software SE (StataCorp. LP, 2013, College Station, Texas, United States). Categorical data were presented using stratified frequency tables (n and %). Trends or associations were assessed using the standard Pearson’s chi-square (χ2) test. If expected cell count in the cross tabulation contained fewer than five observations (sparse numbers), then the Fisher’s exact test was utilised instead. A p-value of less than 0.05 was considered statistically significant.

For the period 2011–2015, a total of 64 502 ESKAPE clinical isolates were recovered from clinical specimens (Figure 1). The distribution of the isolated ESKAPE organisms varied during this period. The clinical specimens included respiratory specimens, blood, urine, catheters, which included central venous and hemodialysis catheters, as well as samples from other sites, including wounds. Respiratory samples included sputum, bronchi alveolar lavage and endotracheal aspirates. Overall, S. aureus (24 495, 38.0%) was the most frequently isolated pathogen followed by K. pneumoniae (14 282, 22.2%), P. aeruginosa (11 231, 17.4%), A. baumannii (8010, 12.4%), Enterobacter spp. (4267, 6.6%), E. faecium (2217, 3.4%) (Table 3 and Figure 1). Similarly, S. aureus (3787, 35.8%) and K. pneumoniae (3059, 28.9%) were the most frequently isolated species from blood cultures. Gram-negative organisms namely K. pneumoniae (2469, 29.2%), A. baumannii (2036, 24.1%) and P. aeruginosa (2015, 23.8%) were the most common isolates from respiratory samples, followed by Gram-positive species, of which S. aureus was the most frequently isolated (1353, 16%). K. pneumoniae, a member of the Enterobacteriaceae group, was the predominant species isolated from urine (4997, 55.5%). Catheters included a combination of central venous, arterial and hemodialysis catheters. A. baumannii (31.5%), S. aureus (23.4%) and K. pneumoniae (23.1%) were the most common isolates cultured. Other sample types included wound swabs, pus swabs and aspirates. The most frequently isolated species from other sample types was S. aureus (18 397, 53.4%).

During the five-year period, high levels of antibiotic resistance were detected among the ESKAPE pathogens. Overall, the most common resistance pattern in Gram-negative ESKAPE pathogens was resistance to amoxicillin-clavulanate and ceftriaxone. The majority of the A. baumannii isolates were resistant to multiple drugs, with over 70% of isolates resistant to all tested antimicrobial agents except colistin, amikacin and gentamicin. Isolates of P. aeruginosa remained fairly susceptible to agents like amikacin, piperacillin-tazobactam, meropenem imipenem and colistin. Overall, colistin resistance was less than 10% over the five-year period for the Gram-negative ESKAPE pathogens.

Between 2011 and 2015, 7788 (54.5%) K. pneumoniae isolates were resistant to amoxicillin-clavulanate. More than 50% of all isolates were resistant to third-generation cephalosporins, including ceftriaxone (8326, 58.3%) and ceftazadime (7412, 51.9%), indicative of ESBL production; 5317 (37.2%) isolates were resistant to ciprofloxacin. Of the 14 282 K. pneumoniae isolates, 922 (6.4%) were resistant to meropenem (Table 4). Increasing antimicrobial resistance trends were observed in K. pneumoniae (Figure 2). Resistance to ceftriaxone increased from 54.6% in 2011 to 65.5% in 2015 (p = 0.018). For the period 2011–2014 average resistance to ciprofloxacin was 35%, which increased to 42% in 2015 (p = 0.015). Resistance to amoxicillin-clavulanate increased from 50.6% in 2011 and to 62.7% in 2015 (p = 0.009). Between 2011 and 2014, resistance to meropenem was 5% or less. This was followed by a significant increase to 16% in 2015 (p < 0.001). Colistin resistance in K. pneumoniae was approximately 2%, but this was not confirmed by reference or molecular methods. Both European Committee on Antimicrobial Susceptibility Testing and the Clinical Laboratory Standards Institute recommend broth microdilution for antimicrobial susceptibility testing of colistin, but broth microdilution is rarely used in routine microbiology laboratories.¹² Broth microdilution is not routinely performed at these KwaZulu-Natal public sector microbiology laboratories and as a result they may have failed to detect heteroresistance.

High MDR rates were observed for A. baumannii with over 70% of isolates resistant to meropenem, imipenem, ciprofloxacin, piperacillin-tazobactam and ceftazidime (Table 4).

Accumulative average resistance to piperacillin-tazobactam, ceftazadime, ciprofloxacin, meropenem, and amikacin observed in P. aeruginosa was 1706 (15.2%), 1601 (14.2%), 1501 (13.4%), 1090 (9.7%) and 716 (6.3%), (p < 0.001) (Table 4). Although 840 (7.5%) samples of P. aeruginosa, were resistant to colistin, this was not confirmed by reference or molecular methods. Over the five-year period, an increase in the total number of P. aeruginosa was noted (p = 0.002) (Figure 1). During 2013–2015, decreasing trends in resistance patterns were observed in P. aeruginosa to ceftazadime (from 17% to 13%, p = 0.004), piperacillin-tazobactam (from 27% to 21%, p < 0.001), meropenem (from 18% to 10%, p < 0.001), ciprofloxacin (from 22% to 18%, p = 0.002), and amikacin (from 10% to 8%, p < 0.001) (Figure 3).

During the five-year period, the average accumulative resistance rate to available carbapenems observed in Enterobacter spp. was 5%. Resistance to ciprofloxacin was greater in E. cloacae (16%) compared to E. aerogenes (8%) (data not shown).

As illustrated in Figure 4, rates of MRSA varied between 18% and 31% per year. Overall, a decreasing trend in the proportion of MRSA was observed between 2011 and 2014, 28% to 18% (p < 0.001). Decreasing resistance rates or trends across several antibiotic classes, namely gentamicin, clindamycin, ciprofloxacin, erythromycin and rifampicin, were also noted. No resistance to vancomycin, teicoplanin and linezolid was observed. Of the 24 495 isolates of S. aureus isolated over the study period, 22 596 (92%) were resistant to penicillin, 12 168 (49%) were resistant to co-trimoxazole, 6034 (24%) were resistant to cloxacillin, 5352 (21.8%) were resistant to ciprofloxacin, 4916 (20%) were resistant to erythromycin, 4702 (19%) were resistant to rifampicin, and 917 (3.7%) were resistant to clindamycin. All isolates of E. faecium (2217) were susceptible to vancomycin and linezolid.

ESBL-producing K. pneumoniae, increased from 54.9 in 2011 to 65.5% in 2015 (p = 0.018). MRSA decreased from 31.4% in 2012 to 19.1% in 2015 (p < 0.001). MDR A. baumanni decreased from 85% in 2012 to 70% in 2015 (Figure 5). Overall, during the five-year period 6351 (79.2%) of all A. baumannii were MDR, 6034 (24.6%) of all S. aureus were MRSA and 8511 (59.5%) of K. pneumoniae were ESBL-producers.

There are a number of limitations of this study. The total number of samples received by all the laboratories during the study period was not available. There were deficiencies in data collection for 2014 due to an upgrade of microbiological diagnostic systems, resulting in unrecoverable surveillance data. Additionally, although this surveillance reflects nine hospitals across two districts in KwaZulu-Natal, the private health sector was not included, and the results reflect only the public health sector in KwaZulu-Natal, South Africa.

This is the first study describing resistance trends in WHO global priority pathogens within KwaZulu-Natal. The high prevalence rates of antimicrobial resistance observed in ESKAPE pathogens signal the need to improve antimicrobial stewardship and infection prevention and control programmes in the region.