Antimicrobial Resistance (AMR) Surveillance on Blood Culture Specimen in Public Hospitals and Clinics - Hospital Authority AMR Data (2020)

Release date: 13 January 2022

Methodology

Surveillance was conducted following the recommendations of WHO GLASS Manual for Early Implementation (2015) with local adaptation as summarised below:

• Six  WHO  priority  organisms,  namely  Escherichia   coli,  Klebsiella  pneumoniae,  Staphylococcus aureus, Salmonella species (spp.), Acinetobacter spp. and Streptococcus pneumoniae, were reported.
• Location of onset were classified as community-onset (CO) (organism isolated from blood specimen collected in non-inpatient services, or within 48 hours after hospital admission) or hospital-onset (HO) (organism isolated from blood specimen collected more than 48 hours after hospital admission). As Salmonella spp. and Streptococcus pneumoniae rarely cause hospital-associated infections, related AST results were interpreted as  “community (undifferentiated) onset”.
• For each surveillance period (one calendar year), only the first result would be reported for each patient per specimen type per organism for the same location of onset.
• AST result being “Intermediate” or “Resistant” was considered as “Non-susceptible”, while AST results derived from less than 10 isolates per calendar year were excluded from analysis.
• To avoid misleading or interference by selection bias, percentages of non-susceptibility derived from less than 70% of total isolates were not reported (or remarked to remind readers to interpret with caution.)
• Positive cultures for the same organism within a 14-day period from the same patient would be regarded as a single episode even if the first was of CO while the second or the rest were of HO (“Day-14” rule).

Where appropriate, AST results in relation to broad-spectrum antimicrobials identified by experts in HA (Big Guns) were examined because of their importance on treating resistant infections. For Salmonella spp. and Streptococcus pneumoniae, since infection caused by these organisms are usually treated by first-line antimicrobials, non-susceptibility percentage of respective first-line antimicrobials were reported.

Since the Action Plan was published in 2017, the situation of 2016 has been chosen as the baseline for comparison.  Fisher’s exact test or chi-square test was used to compare non-susceptibility percentages^[1] between year 2019 and 2020, whereas One-way Cochran-Armitage test was used to look for trend from year 2016 to 2020.  P-value of less than 0.05 was considered as statistically significant.

Results

In 2020, a total of about 140,000 patients had blood culture collected, showing an decrease of about 10,000 when compared with year 2016. The reduction was contributed mostly by the age group of 0-14 (44.9% drop from 28,000 in 2019 to 16,000 in 2020). Amongst patients with blood culture collected each year, over 50% were aged 65 years or above.  The percentage of patients with positive blood culture remained stable from year 2016 to 2020 (around 10-11%).

The overall distribution of the six WHO priority organisms cultured from blood were similar during 2016 to 2020, with the three commonest organisms in 2020 being Escherichia coli, Klebsiella pneumoniae and Staphylococcus aureus (isolated from 41.3%, 11.9% and 11.2% of patients with positive blood culture respectively).  A significant drop in number of patients with Streptococcus pneumoniae isolated from blood in year 2020 was observed (from about 130 in 2019 to about 30 in 2020). Regarding location of onset, Escherichia coli, Klebsiella spp., Salmonella spp. and Streptococcus pneumoniae were predominantly CO, while Acinetobacter spp. was predominantly HO.

In general, non-susceptibility on majority of selected antimicrobials for the WHO priority organisms remained stable from year 2016 to 2020.  However, increasing trends of non-susceptibility percentage were also observed among several pathogen-antimicrobial combinations that may warrant further monitoring.

Escherichia coli

In general, non-susceptibility percentages were lower among E. coli isolates of CO than those of HO. Comparing between year 2019 and 2020, E. coli isolates showed reduction in non-susceptibility percentage towards cefepime (CO: 18.2%→15.1%; HO: 25.2%→19.9%) which was statistically significant.  Statistically significant decreasing trends in non-susceptibility percentages from year 2016 to 2020 were also observed towards piperacillin/tazobactam (CO: 7.2%→ 3.8%; HO: 14.1%→10.7%), ceftazidime (HO: 23.2%→17.4%) and cefepime (CO: 23.6%→15.1%; HO: 33.1%→19.9%).On the contrary, statistically significant increasing trend from year 2016 to 2020 towards imipenem (CO: 0.04%→0.1%^[3]) was observed.

Klebsiella pneumoniae

Non-susceptibility percentages were generally lower among CO isolates than those of HO.  No statistically significant changes in non-susceptibility was observed towards broad-spectrum antimicrobials tested among CO and HO isolates between year 2019 and 2020.  Trend analysis revealed statistically significant increasing trends from year 2016 to 2020 towards ceftazidime (CO: 7.3%→9.9%), meropenem (HO: 1.2%^[4]→4.6%) and imipenem (HO: 1.3%→4.0%^[5]), while statistically significant decreasing trend was observed for piperacillin/tazobactam (HO: 21.5%→15.7%) and cefepime (CO: 8.4%→5.5%).

Staphylococcus aureus

Non-susceptibility percentage of oxacillin^[6] for Staphylococcus aureus of HO were generally higher than those of CO.  In 2020, 0.1% (1 out of 678 isolates) of the Staphylococcus aureus isolates of HO were found non-susceptible towards vancomycin.

Salmonella species

Neither statistically significant change nor trend were observed among broad-spectrum antimicrobials for Salmonella spp. isolates between year 2019 and 2020 or from year 2016 to 2020.  Trend analysis revealed statistically significant increasing trend from year 2016 to 2020 towards ampicillin (62.4%→74.5%), while statistically significant decreasing trend was observed for ceftriaxone (6.2%→2.2%)

Acinetobacter species

Non-susceptibility percentages were lower among Acinetobacter spp. isolates of CO than those of HO. By comparing year 2019 and 2020, statistically significant increasing non-susceptibility percentages were observed towards piperacillin/tazobactam (HO: 58.3%→73.5%), cefoperazone/sulbactam (HO: 51.8%→65.7%), cefepime (HO: 58.6%→73.9%), meropenem (HO: 66.3%^[7]→80.2%) and imipenem (HO: 54.1%→74.0%). Statistically significant increasing trends from year 2016 to 2020 towards meropenem (HO: 59.0%^[8]→80.2%) and imipenem (HO: 60.2%→74.0%) were observed.

Streptococcus pneumoniae

Non-susceptibility percentage of co-trimoxazole and erythromycin for Streptococcus pneumoniae isolates remained above 50% from year 2016 to 2020.  While there was no statistically significant change identified among selected antimicrobials for S. pneumoniae isolates between year 2019 and 2020, statistically significant increasing trends during 2016 to 2020 were noted for non-susceptibility percentages towards penicillin (0.7%→14.3%) and cefotaxime (0.9%→13.6%).

Trends in non-susceptibility percentages from year 2016 to 2020 with statistical significance mentioned above were summarised in the tables below.

Recommendations

In view of statistically significant increasing trend of non-susceptibility for HO isolates of Klebsiella pneumoniae spp. and Acinetobacter spp. towards meropenem and imipenem, further monitoring would be warranted.  We have alerted working partners of HA about increasing trend of non-susceptibility of the aforesaid broad-spectrum antimicrobial-organism combinations for their further investigation and management as appropriate.

Please click here to view further details.

Acknowledgement

We would like to acknowledge stakeholders from HA such as Information Technology & Health Informatics Division, Quality and Safety Division, Strategy and Planning Division, and various Working Groups for provision of input to facilitate compilation of findings for year 2020.

^[1]  Non-susceptibility percentage means the proportion of isolates tested for susceptibility of a particular antimicrobial, and with test result being non-susceptible.

^[2]  The following section illustrates some important drug-bug combinations.  Please refer to the powerpoint for full results. Data related to levofloxacin should be interpreted with caution because CLSI guidelines for sensitivity testing involving levofloxacin interpretive criteria for Enterobacteriaceae (except Salmonella spp.) has been updated in 2019.  For laboratories that chose to apply the new criteria for reporting in 2019, some E. coli and K. pneumoniae isolates previously categorised as susceptible to levofloxacin would be categorised as non-susceptible using the updated zone size requirement under the 2019 criteria.

^[3]  Non-susceptibility percentage should be interpreted with caution as the figure is derived from less than 70% of total Escherichia coli isolates for surveillance. The figure may be affected by selection bias.

^[4]  Non-susceptibility percentage should be interpreted with caution as the figure is derived from less than 70% of total K. pneumoniae isolated for surveillance. The figure may be affected by selection bias..

^[5]  Non-susceptibility percentage should be interpreted with caution as the figure is derived from less than 70% of total K. pneumoniae isolated for surveillance. The figure may be affected by selection bias.

^[6]  Sensitivity testing results of penicillinase stable penicillins (oxacillin, cloxacillin and methicillin) and cefoxitin towards Staphylococcus aureus were collectively grouped as “oxacillin” following the recommendation of CLSI.

^[7]  Non-susceptibility percentage should be interpreted with caution as the figure is derived from less than 70% of total Acinetobacter spp. isolates for surveillance. The figure may be affected by selection bias.

^[8]  Non-susceptibility percentage should be interpreted with caution as the figure is derived from less than 70% of total Acinetobacter spp. isolates for surveillance. The figure may be affected by selection bias.