Summary of Medical PhD thesis: Investigation of beta-lactam resistance and genes encoding beta-lactamase of escherichia coli and klebsiella pneumoniae causing bloodstream infections

MINISTRY OF EDUCATION AND TRAINING MINISTRY OF DEFENSE

108 INSTITUTE OF CLINICAL MEDICAL AND PHARMACEUTICAL SCIENCES

TRINH VAN SON

INVESTIGATION OF BETA-LACTAM RESISTANCE

AND GENES ENCODING BETA-LACTAMASE OF

ESCHERICHIA COLI AND KLEBSIELLA PNEUMONIAE

CAUSING BLOODSTREAM INFECTIONS

Specialty: Infectious and Tropical Diseases

Educational code: 62720153

SUMMARY OF MEDICAL PhD THESIS

Hanoi - 2021

THE RESEARCH WAS DONE AT:

108 INSTITUTE OF CLINICAL MEDICAL AND PHARMACEUTICAL SCIENCES

Supervisors:

1. A/Prof. Le Huu Song, MD, PhD

2. Dr. Nguyên Đang Manh, MD, PhD

Reviewer 1: ……. Reviewer 2: ……. Reviewer 3: …….

The Institutional thesis defense meeting will take place at 108 Institute of Clinical Medical and Pharmaceutical Sciences At …… h …. month day, year

The thesis was stored at:

1. National library of Vietnam 2. The Library of 108 Institute of Clinical Medical and

Pharmaceutical Sciences

INTRODUCTION

Significance of the research

Bloodstream infection (BSI) is a potentially life-threatening

condition and has become a major challenge for healthcare system

worldwide, that have affected the million people with increasing

morbidity and mortality, especially BSI caused by multidrug resistant

bacteria. The reports demonstrate the percentage of bacteria producing

extended spectrum beta-lactamase (ESBL) are increasing from 39% to

55% in Southeast Asia. Vietnam is one of the countries which have the

highest prevalence of Enterobacteriaceae producing ESBL (55.1%).

Previous studies have reported that the prevalence of carbapenem

resistant Enterobacteriaceae in Vietnam was from 5.6% to 20%.

Research on antibiotic resistance genes is mainly used in

investigating the epidemiology as well as transmission mechanism.

Some studies demonstrated the ability to use evidences of genetic

resistance to guide empirical antibiotic therapy in clinical practice. In

Vietnam, the infections caused by Cephalosporin and Carbapenem

resistant Enterobacteriaceae have been a major burden. However,

national data of genetic resistance encoding ESBL and carbapenemase

are insufficient. Therefore, the aims of this study were:

1. Evaluating characteristics of beta-lactam resistance and

distribution of genes encoding beta-lactamase of Escherichia coli and

Klebsiella pneumoniae causing bloodstream infections at 108 Military

Central Hospital (from October 2014 to march 2016).

2. Assessing the value of genes encoding beta-lactamase to

identify phenotypic beta-lactam resistance of bacteria and influence

on treatment response.

MAIN SCIENTIFIC CONTRIBUTION

1. High occurrence of broad-spectrum cephalosporin

resistance of E. coli and K. pneumoniae causing BSI were reported.

The prevalence of ESBL-producing E. coli and K. pneumoniae were

58.3% and 22,0% respectively. 26,0 % of K. pneumoniae strains

resisted to carbapenem.

2. There were 91.3% of E. coli and 90% of K. pneumoniae

carrying at least one of three common genes encoding ESBL (CTX-

M, TEM and SHV).

3. The CTX-M gene is valuable in diagnosis of cephalosporin

resistance of E. coli and K. pneumoniae with the sensitivity of 77.8%-

90.2%. The specificity of NDM-1 to identify carbapenem resistant K.

pneumoniae was 86,5%.

4. In the group of BSI patients caused by cephalosporin

susceptible strains, the treatment response was worse in BSI patients

due to strains carrying CTX-M.

5. The findings of study provide further clinical evidence for

the use of antibiotic resistant genes to guide antibiotic treatment

decisions in patients with BSI caused by E. coli and K. pneumoniae.

STRUCTURE OF THE THESIS

The thesis includes 116 pages, with four chapters, introduction 02 pages, chapter 1 - Literature review 34 pages, chapter 2 - Materials and Methods 21 pages, chapter 3 - Results 27 pages, chapter 4 - Discussion 30 pages, conclusions and proposal 02 pages. The thesis has 41 tables, 07 pictures, 12 figures, 04 diagrams, 194 references with 12 in Vietnamese and 172 in English.

LITERATURE REVIEW

Epidemiology and etiology of bloodstream infection

Bloodstream infection (BSI) have been being a challenge for

healthcare system over the world with high morbidity and mortality.

There are approximately 19 million BSI patients per year with 5

million deaths in the developing and poor countries. In developed

countries, such as United State have about 970000 BSI patients per

year and numbers of patients are continuously increasing.

In recent years, the data on etiology of BSI demonstrated that

the bacterial Gram negative was the most common pathogen, including

community and hospital acquired infections. Namely, E. coli and K.

pneumoniae were the main recorded pathogens. The multiple central

(SENTRY) in 2019 show that S. aureus, E. coli and K. pneumoniae

were the most common pathogens causing BSI. In detail, E. coli was

increasing from 18.5% (1997-2000) to 24.0% (2013-2016). Some

reports from Vietnam and other countries in Southeast Asia showed

the similar results.

Enterobacteriaceae producing beta-lactamase

1.2.1. Molecular mechanisms of beta-lactamase

Enterobacteriaceae is a large family of Gram-negative

bacteria, including a lot of important human pathogens. The common

mechanism of beta-lactam resistance is producing beta-lactamases,

mostly ESBL and carbapenemase. ESBLs are enzymes to hydrolyze

broad spectrum beta-lactam, including the most of penicillin,

cephalosporin (except for cephamycin) and monobactam. The most

common genes encoding ESBL belong to class A (as following

Ampler classification) are TEM, SHV and CTX-M. Gram-negative

bacteria producing carbapenemases are major concern for clinical

practice because of resistance to broad spectrum of beta-lactam

substrates, including carbapenems. The family of carbapenemases is

diverse, including class A (SME, IMI, GES and KPC), class B (VIM,

IMP, NDM) and class D (OXA-23, OXA-48, OXA-51).

1.2.2. Molecular epidemiology

The reports show that beta-lactamases are encoded by lots of

genes on plasmid or chromosome. Therefore, they are easy to transmit

between species of Enterobacteriaceae family. The common gene family

encoding ESBL are CTX-M, SHV and TEM. The molecular

epidemiology research performed the proportion of E. coli and K.

pneumoniae carrying CTX-M was about from 66.7% to 99.4%. Namely,

the most common sub genotype was CTX-M-14, CTX-M-15. Other

genes were reported, those were OXA-1, TEM-1 and SHV-1. The reports

from Europe and United States about carbapenemase producing

Enterobacteriaceae showed that most common genes were KPC (42%-

69.9%) and OXA-48 (38%) and rarer genes were NDM and VIM.

However, in Pacific Asia, NDM-carbapenemase and other metallo-beta-

lactamases (IMP, VIM) are more common than KPC-beta-lactamase,

especially in Southeast Asia.

In Vietnam, the reports about Enterobacteriaceae producing

ESBL in period 2007-2013 demonstrated that prevalence of TEM, SHV,

CTX-M was 83.5%-87.8%, 44.3%-62.2%, 24.6%-36%, respectively. A

study (2017) about strains producing ESBL causing BSI, the most

common gen was CTX-M (95%), following up TEM (45%), OXA (20%)

and SHV (3%). Studies on molecular mechanisms showed that NDM was

a common gene identified from carbapenem-resistant strains.

Techniques to identify bacterial pathogens and antibiotic

susceptibility testing in bloodstream infections

So far, blood culture is still a “gold standard” to identify

pathogens for BSI. Nowadays, blood culture protocol in clinical

practice is used automated platform to get real-time results. The culture

medium is continuous improvement to increase spectrum bacteria

culture and also reduce time for culture. However, blood culture

methods have some limitations as low sensitivity, high blood volume

imperfect growth for difficulty to culture pathogens and time-

consuming matter. Antibiotic susceptibility testing (AST) is tested

with different methods for each microbiology lab, such as disk

diffusion or minimal inhibitory concentration (MIC). In clinical

practice, AST is done by automated system such as Vitek or Phoenix

system based on MIC method.

Other methods to identify pathogens base on polymerase

chains reaction (PCR) tecnique to detect bacterial DNA in clinical

samples. In BSIs, the methods may use to indirectly identify bacteria

from positive cultures or directly from clinical samples (blood,

sputum, cerebral spinal fluid, urine). Multiplex-PCR for identification

of a broad range of pathogens are developing. However, the methods

are not accurately consistant when compared with blood culture. In

additional, the methods can detect antibioic resistant genes. Moreover,

until now genotypic resistance are not enough evidence to assess

phenotypic drug resistance and the methods are also so expensive to

apply in clinical practice.

MATERIALS AND METHODS

Materials

2.1.1. Study population

165 BSI patients were included in this study. They were

diagnosed and treated at the 108 Military Central Hospital from Oct,

2014 to May, 2016. In which, 115 patients caused by E. coli and 50

patients caused by K. pneumoniae.

2.1.2. Inclusion criteria

Bloodstream infections were diagnosed with criteria following

Surviving sepsis campaign: international guidelines for management

of severe sepsis and septic shock: 2012 and documented infections

were positive blood culture with E. coli or K. pneumoniae.

2.1.3. Exclusion criteria

It did not match the results of bacteria between culture bottles or

polyinfection. Patients or their relatives did not agree to join the study.

Methods

2.2.1. Study design: Observational study.

2.2.2. Research parameters

2.2.2.1. Clinical and subclinical parameters

Clinical and subclinical characteristics of BSI patients were

recorded with the same study report for each patient. Clinical variables

were examined by physicians and subclinical variables were tested

with routine tests in the hospital.

2.2.2.2. Bacterial culture and antibiotic susceptibility testing

Blood cultures were collected when patients admitted with

community infections or started onset of BSI with hospital acquired

infections. Bacterial identification and antibiotic susceptibility testing

were done with Vitek II Compact System (BioMerieux, France).

Antibiotic substances were tested including Ampicillin,

Amocixillin/Acid Clavulanic (AMC), Piperacillin/Tazobacbam

(TZP), Cefotaxime (CTX), Ceftazidime (CAZ), Cefepime (FEP),

Ertapenem (ETP), Imipenem (IPM), Meropenem (MEM).

2.2.2.3. Genes encoding beta-lactamases

Genes encoding beta-lactamases were investigated including:

six genes encoding ESBL (TEM, SHV, CTX-M, VEB, PER, GES) and

nine genes encoding carbapenemases (KPC, AIM, IMP, VIM, SPM,

NDM-1, OXA-58, OXA-23 and OXA-48). Combination of genes

encoding beta-lactamse was also reported such as 3-POS, that was

strains carrying all of three genes CTX-M, TEM and SHV.

2.2.2.4. Treatment response: evaluated by existence of septic shock,

fatality rate, duration of fever after diagnosis and length of hospital stay.

Statistical analysis

Statistical analyses were performed using the SPSS software

v.23.0 (IBM Corporation, USA). Continuous variables are presented

as mean ± standard deviation. Categorical variables are given as

frequencies with percentages and comparisons of categorical variables

between groups were performed using chi-square and Fisher’s exact

tests. The level of significance was set at p-values < 0.05.

Ethics statements

Ethical approval was obtained from the Ethics Committee and

the Review Board of the 108 Military Central Hospital. The Ethics

Committee also approved a waiver of informed consent to participate

in this study because of its retrospective design. All patient data was

anonymized prior to the analysis.

RESULTS

General characteristics

Table 3.1. Baseline characteristics of patient cohort

Characteristics

Sex (male) n (%)

E. coli BSIs (n=115) 70(60.9)

K. pneumoniae BSIs (n=50) 37(74.0)

Total (n=165) 107(64.8)

Mean of age (years)

62.3±16.2

62.0±17.2

62.3±16.5

62(53.9)

26(52.0)

88(53.3)

 60 years old n (%)

73(63.5)

27(54.0)

100(60.6)

Pre-existing conditions n (%)

44(88.0)

139(84.2)

Primary source infection n (%)

95(82.6)

Mean of SOFA (points)

3.36±3.05

3.62±3.1

3.44±3.06

While blood cell (G/L)

16.9±13,7

16.4±8.9

16.8±12.4

Neutrophile (G/L)

13.7±9.9

13.9±8.0

13.8±9.3

Platelet (G/L)

190±118

170±130

184±122

Prothrombin (%)

76.5±22.8

81.7±21.6

77.7±22.5

Pro-calcitonin (ng/L)

32.53±38.23

29.54±32.49

31.58±36.41

Lactate (mmol/L)

5.34±3.53

4.89±3.18

5.14±3.34

5.2±2.5

6.7±3.8

5.6±3.0

Duration of fever after diagnosis (days) Length of hospital stay (days)

19.9±14.7

27.7±27.6

22.3±19.8

Septic shock n (%)

19(16.5)

14(28.0)

33(20.0%)

Mortality n (%)

18(15.7)

12(24.0)

30(18.2%)

The results show that the most patients were male (64.8%) and

over 60 years old (53.3%) with pre-existing conditions. The primary

source infections were reported 82.6% with E. coli-BSIs and 88.0%

with K. pneumoniae-BSIs. Subclinical of BSI patients performed

increasing while blood cell with the most Neutrophil, pro-calcitonin

and serum lactate. Mean of SOFA was higher than 3 points. There

were not different significance subclinical variables between E. coli-

BSIs and K. pneumoniae-BSIs. The ratio of septic shock and mortality

were 20.0% and 18.2% respectively.

The status of beta-lactam resistance and characteristic of

genes encoding beta-lactamases

3.2.1. The status of beta-lactam resistance

Table 3.5. The percentage of cephalosporin resistance

Strains

E. coli (n=115)

K. pneumoniae (n=50)

p*

Antibiotic

58.3

Positive

0.001

22.0

ESBL

41.7

Negative

78.0

0.919

4.3

Susceptible

4.0

0.9

Intermediate

12.0

AM

94.8

Resistant

109

84.0

0.091

58.3

Susceptible

44.0

13.9

Intermediate

6.0

AMC

27.8

Resistant

50.0

0.001

80.9

Susceptible

52.0

12.2

Intermediate

4.0

TZP

7.0

Resistant

44.0

0.006

29.6

Susceptible

52.0

0.0

Intermediate

0.0

CTX

70.4

Resistant

48.0

0.350

53.9

Susceptible

46.0

0.9

Intermediate

0.0

CAZ

45.2

Resistant

54.0

0.815

62.4

Susceptible

60.4

11.9

Intermediate

10.4

FEP**

Resistant

29.2

25.7 Note: (*) comparison susceptible strains between E. coli and K. pneumoniae; (**) 157 strains were tested with FEP, including 109 E. coli and 48 K. pneumoniae.

E. coli strains were high resistance to AM (94.8%) and CTX

(70.4%). However, they were still susceptible to FEP (62.4%) and

AMC (58.3%). Similarly, K. pneumoniae strains were high resistance

to AM (84.0%) but still susceptible to FEP (60.4%). The proportions

of TZP susceptible E. coli were significantly higher than K.

pneumoniae (80.9% vs 52.0% with p = 0.001).

Table 3.6. The percentage of Carbapenem resistance

Strains

E. coli (n=115)

K. pneumoniae (n=50)

p*

Antibiotics

0.001

97.4

Susceptible

112

74.0

1.7

ETP

Intermediate

0.0

0.9

Resistant

26.0

0.001

100

Susceptible

115

74.0

0.0

IPM

Intermediate

2.0

0.0

Resistant

24.0

0.001

100

Susceptible

115

74.0

0.0

MEM

Intermediate

0.0

Resistant

0.0

26.0 Note: (*) comparison susceptible strains between E. coli and K. pneumoniae.

There were three E. coli strains, they were not susceptible to

ETP but still susceptible to IPM and MEM. The proportions of

Carbapenem susceptible K. pneumoniae was 13/50 (26.0%).

42/50

80/115

67/115

23/50

5/50

1/115

0/115

90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0%

E. coli

TEM 58.3%

SHV 0.9%

CTX-M 69.6%

VEB 0.0%

K. pneumoniae

46.0%

84.0%

46.0%

10.0%

3.2.2. Characteristics of genes encoding beta-lactamase

Figure 3.1. Contributions of gene encoding ESBL

All of genes were tested: E. coli strains were the most detected

with CTX-M and TEM. The similar to K. pneumoniae were SHV;

CTX-M and TEM. PER and GES were not detected in any strains.

12/50

25.0%

20.0%

6/50

15.0%

9/115

10.0%

5/115

1/50

5.0%

0/115

0.0%

E. coli

NDM-1 7.8%

VIM 4.3%

KPC 0.0%

OXA-48 0.0%

K. pneumoniae

24.0%

12.0%

2.0%

Figure 3.2. Contributions of genes encoding carbapenemase

There were 12/50 (24%) of K. pneumoniae carrying NDM-1,

that was higher than E. coli 9/115(7.8%). 5/115 (4,3%) of E. coli and

105/115

45/50

24/50

21/50

20/50

19/50

43/115

42/115

18%

1/115

0/115

00%

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 00%

TEM+SHV

1-POS

2-POS

3-POS

TSCN

E. coli

00%

91%

37%

00%

TEM+ CTX-M 37%

SHV+ CTX-M 01%

K. pneumoniae

40%

90%

48%

38%

18%

42%

6/50 (12%) of K. pneumoniae were detected with VIM.

Figure 3.3. Contributions of strains carrying combination genes

High prevalence of strains was detected with at least one of

three gen (1-POS), 91.3% with E. coli and 90% with K. pneumoniae.

There were 38% of K. pneumoniae carrying all of three genes (TEM,

SHV and CTX-M) and 9/50 (18.0%) of K. pneumoniae carrying all of

4 genes (TEM; SHV; CTX-M and NDM-1). However, there were not

any E. coli strains carrying 3-POS or TSCN.

Correlation between genotypic resistance and beta-lactam

resistance and influence on the outcome of BSI patients.

3.3.1. Correlation between genotypic resistance and phenotypic beta-

lactam resistance.

Table 3.7. Correlation between genotype and phenotype of Penicillin and Cephalosporin resistance of E. coli

Genes

Antibiotics

Positive (n=67)

ESBL

TC n (%) 31 (46.3) 11 (22.9)

TEM n (%) 38 (56.7) 29 (60.4)

CTX-M n (%) 58 (86.6) 22 (45.8)

1-POS n (%) 65 (97.0) 40 (83.3)

2-POS n (%) 31 (46.3) 12 (25.0)

NoT n (%) 40 (60.4) 29 (59.7)

NoC n (%) 60 (89.6) 22 (45.8)

0.010 37 (45.7)

0.692 46 (56.8)

0.001 69 (85.2)

0.010 78 (96.3)

0.020 37 (45.7)

0.938 48 (59.3)

0.001 71 (87.7)

Negative (n=48) p Non-Sus (n=81)

CTX

5 (14.7)

21 (61.8)

11 (32.4)

27 (79.4)

6 (17.6)

21 (61.8)

11 (32.4)

0.002 22 (41.5)

0.622 28 (52.8)

0.001 45 (84.9)

0.003 51 (96.2)

0.005 23 (43.4)

0.802 29 (54.7)

0.001 46 (86.8)

Sus (n=34) p Non-Sus (n=53)

CAZ

20 (32.3)

39 (62.9)

35 (56.5)

54 (87.1)

20 (32.3)

40 (64.5)

36 (58.1)

0.304 18 (43.9)

0.275 21 (51.2)

0.001 37 (90.2)

0.083 40 (97.6)

0.218 19 (46.3)

0.285 22 (53.7)

0.001 38 (92.7)

Sus (n=62) p Non-Sus (n=41)

FEP

20 (29.4)

42 (61.8)

37 (54.4)

59 (86.8)

20 (29.4)

43 (63.2)

38 (55.9)

0.124 20 (41.7)

0.280 35 (72.9)

0.001 32 (66.7)

0.059 47 (97.9)

0.074 20 (41.7)

0.324 36 (75.0)

0.001 34 (70.8)

Sus (n=68) p Non-Sus (n=48)

AMC

22 (32.8)

32 (47.8)

48 (71.6)

58 (86.6)

22 (34.3)

33 (49.3)

48 (71.6)

Sus (n=67) p Non-Sus (n=22)

TZP

0.332 10 (45.5) 32 (34.4)

0.007 15 (68.2) 52 (55.9)

0.567 16 (72.7) 64 (68.8)

0.033 21 (95.5) 84 (90.3)

0.423 10 (45.5) 33 (35.5)

0.005 15 (68.2) 54 (58.1)

0.925 16 (72.7) 66 (71.0)

0.333

0.294

0.720

0.442

0.385

0.384

0.870 Sus (n=93) p

In all of genes encoding beta-lactamases of E. coli strains,

CTX-M and NoC were detected significant higher in Cephalosporin

resistant E. coli than the rest group with p = 0.001. TEM-E. coli was

significant higher in AMC non-susceptible E. coli than susceptible

group with p=0.007.

Table 3.10. Correlation between genotype and phenotype of Penicillin and Cephalosporin resistance of K. pneumoniae

Genes

Antibiotics

CTX-M n (%) 10 (90.9)

TEM n (%) 10 (90.9)

SHV n (%) 9 (81.8)

TC n (%) 9 (81.8)

NoT n (%) 10 (90.9)

NoC n (%) 11 (100)

NoS n (%) 11 (100)

Positive (n=11)

ESBL

13 (33.3)

33 (84.6)

12 (30.8)

14 (35.9)

33 (84.6)

Negative (n=39)

0.001 p

0.001

0.823

0.002

0.001

0.166

19 (79.2)

18 (75.0)

20 (83.3)

17 (70.8)

19 (79.2)

21 (87.5)

22 (91.7)

Non-Sus (n=24)

CTX

4 (15.4)

5 (19.2)

22 (84.6)

4 (15.4)

5 (19.2)

4 (15.4)

22 (84.6)

Sus (n=26)

0.001 p

0.001

0.902

0.001

0.443

21 (77.8)

20 (74.1)

23 (85.2)

19 (70.4)

21 (77.8)

23 (85.2)

25 (92.6)

Non-Sus (n=27)

2 (8.7)

3 (13.0)

19 (82.6)

CAZ

2 (8.7)

3 (13.0)

2 (8.7)

19 (82.6)

Sus (n=23)

0.001 p

0.001

0.804

0.001

0.279

17 (89.5)

15 (78.9)

17 (89.5)

15 (78.9)

17 (89.5)

Non-Sus (n=19)

FEP

5 (17.2)

7 (24.1)

23 (79.3)

5 (17.2)

7 (24.1)

6 (20.7)

25 (86.2)

Sus (n=29)

0.001 p

0.001

0.356

0.001

0.738

22 (78.6)

20 (71.4)

25 (89.3)

20 (71.4)

21 (75.0)

23 (82.1)

26 (92.9)

Non-Sus (n=28)

AMC

1 (4.5)

3 (13.6)

17 (77.3)

1 (4.5)

3 (13.6)

2 (9.1)

18 (81.8)

Sus (n=22)

0.001 p

0.001

0.250

0.001

0.233

20 (83.3)

19 (79.2)

21 (87.5)

19 (79.2)

20 (83.3)

21 (87.5)

22 (91.7)

Non-Sus (n=24)

TZP

3 (11.5)

4 (15.4)

21 (80.8)

2 (7.7)

4 (15.4)

22 (84.6)

Sus (n=26)

p

0.001

0.517

0.001

0.443

In all of genes encoding beta-lactamase, TEM, CTX-M, NoC

and NoT were the most contribution of Cephalosporin resistant K.

pneumoniae. The distribution of SHV between susceptible and non-

susceptible cephalosporin group were not different significance.

Table 3.13. Correlation between genotype and phenotype of Carbapenem resistance of K. pneumoniae

Carbapenem

Non- Susceptible (n=13)

Susceptible (n=37)

Genotypes

7 (53.8)

5 (13.5)

0.003

NDM-1 n (%)

0 (0)

6 (16.2)

0.122

VIM n (%)

8 (61.5)

11 (29.7)

0.042

3-POS n (%)

10 (76.9)

14 (37.8)

0.015

2-POS n (%)

6 (46.2)

3 (8.1)

0.002

TSCN n (%)

6 (46.2)

4 (10.8)

0.006

CN n (%)

11 (84.6)

14 (37.8)

0.004

NoC n (%)

10 (76.9)

14 (37.8)

0.015

NoT n (%)

11 (84.6)

33 (89.2)

0.662

NoS n (%)

There were 13 carbapenem non-susceptible K. pneumoniae in

our results. The percentage of NDM-1; 3-POS; TSCN and CN were

7/13 (53.8%); 8/13 (61.5%); 6/13 (46.2%) and 6/13 (46.2%)

respectively. Moreover, NDM-1 was detected in carbapenem non-

susceptible K. pneumoniae strains was signifficant higher than

carbapenem susceptible K. pneumoniae strains.

3.3.2. The value of genes encoding beta-lactamase to predict beta-

lactam resistance

Table 3.15. The value of CTX-M and NoC to predict beta-lactam

resistant E. coli Spe (%)

Sen (%)

PPV (%) NPV (%) Accu (%)

Value

ESBL

CTX

CAZ

FEP

AMC

TZP

CTX-M NoC CTX-M NoC CTX-M NoC CTX-M NoC CTX-M NoC CTX-M NoC

86.6 89.6 85.2 87.7 84.9 86.4 90.2 92.7 66.7 70.8 72.7 72.7

54.2 54.2 67.6 67.6 43.5 41.9 45.6 44.1 28.4 28.4 31.2 29.0

72.5 73.2 86.3 86.6 56.3 56.1 50.0 50.0 40.0 41.5 20.0 19.5

74.3 78.8 65.7 69.7 77.1 78.8 88.6 90.9 54.3 57.6 82.9 81.8

73.0 74.8 80.0 81.7 62.6 62.6 62.4 62.4 44.3 46.1 39.1 37.4

The value of CTX-M to predict cephalosporin resistance was

84.9%-90.2% sensitivity; however, the specificity was from 43.5% to

67.6%. When combination between NDM-1 and CTX-M, they were

increasing sensitivity but not specificity to predict cephalosporin

resistance.

Table 3.16. The value of genes to predict cephalosporin resistant

K. pneumoniae Spe (%)

Sen (%)

PPV (%)

NPV (%)

Accu (%)

Value

66.7

TEM

90.9

43.5

96.3

72.0

66.7

CTX-M

90.9

43.5

96.3

72.0

ESBL

64.1

NoT

90.9

41.7

96.2

70.0

64.1

NoC

100

44.0

100

72.0

80.8

TEM

78.3

77.8

78.0

75.0

84.6

CTX-M

82.6

81.5

82.0

79.2

CTX

80.8

NoT

79.2

80.8

80.0

79.2

84.6

NoC

84.0

88.0

86.0

87.5

87.0

TEM

87.0

74.1

80.0

74.1

91.3

CTX-M

91.3

77.8

84.0

77.8

CAZ

87.0

NoT

87.5

76.9

82.0

77.8

91.3

NoC

92.0

84.0

88.0

75.9

TEM

85.2 78.9

68.2

84.6

77.1

82.8

CTX-M

89.5

77.3

92.3

85.4

FEP

75.9

NoT

78.9

68.2

84.6

77.1

79.3

NoC

89.5

73.9

92.0

83.3

TEM and CTX-M were potential prediction cephalosporin

resistant K. pneumoniae. The sensitivity was from 74.1% to 90.9% and

the specificity was from 66.7% to 91.3%. Combination NDM-1 with

TEM or CTX-M, that made increased sensitivity to predict

cephalosporin resistance but not specificity.

Table 3.18. The value of genotypic resistance to predict carbapenem resistant K. pneumoniae

Value

Sen (%)

Spe (%)

PPV (%)

NPV (%)

Accu (%)

NDM-1

53.8

86.5

58.3

84.2

78.0

3-POS TSCN CN NoC

61.5 46.2 46.2 84.6

70.3 91.9 89.2 62.2

42.1 66.7 60.0 44.0

83.9 82.9 82.5 92.0

68.0 80.0 78.0 68.0

NDM-1 gene is a good candidate to predict carbapenem

resistant K. pneumoniae with 86.5% specificity. Strains carrying all of

four genes (TEM, SHV, CTX-M and NDM-1) had high specificity

(91.9%) but moderate sensitivity (46.2%).

p=0.473

p=0.205

p=0.399

p=0.236

5 3 / 9

2 4 / 1 1

5 1 / 5

7 3 / 8

2 5 / 0 1

5 4 / 8

5 5 / 9

7 4 / 7

ESBL negative

Susceptible to CTX

Susceptible to CAZ

Susceptible to FEP

CTX-M positive

25.7%

33.3%

21.6%

26.2%

19.2%

16.4%

14.6%

3.3.3. Relation between genotypic resistance and clinical outcome

CTX-M negative 17.8% Figure 3.4. Contribution of CTX-M on septic shock patients caused by susceptible cephalosporin strains

p=0.264

p=0.253

p=0.230

5 1 / 4

p=0.378

5 3 / 8

2 4 / 9

7 3 / 7

2 5 / 8

5 4 / 6

5 5 / 7

8 4 / 5

ESBL negative

Susceptible to CTX

Susceptible to CAZ

Susceptible to FEP

CTX-M positive

22.9%

26.7%

18.9%

21.4%

CTX-M negative

15.4%

13.3%

10.4%

12.7%

Figure 3.5. Contribution of CTX-M on mortality patients caused by susceptible cephalosporin strains

In the BSI patients caused by cephalosporin susceptible

strains, the proportion of septic shock (fig 3.6) and mortality (fig 3.7)

were higher in group with CTX-M positive than group with CTX-M

negative. However, the difference was not significant with p > 0.05.

Table 3.19. Duration of fever after diagnosis in BSI patients caused by cephalosporin susceptible strains

Duration of fever after diagnosis (days) (Mean ± SD (n))

Phenotypic resistance ESBL negative Susceptible to CTX Susceptible to CAZ Susceptible to FEP

CTX-M (+) 7.0±4.6 (n=27) 7.8±6.2 (n=11) 6.1±4.2 (n=30) 6.0±4.0 (n=33)

CTX-M (-) 5.2±2.0 (n=44) 15.0±1.9 (n=39) 4.9±2.0 (n=43) 5.0±2.0 (n=48)

p 0.027 0.016 0.094 0.138

In BSI group caused by cephalosporin susceptible strains, the

duration of fever after diagnosis was significant longer in group with

CTX-M positive than group with CTX-M negative.

Table 3.1. The length of hospital stays in septic patients caused by cephalosporin susceptible strains

Length of hospital stay (days) (Mean ± SD (n))

Phenotypic resistance

ESBL negative Susceptible to CTX

CTX-M positive 31.9±29.1 (n=35) 26.4±16.8 (n=15)

CTX-M negative 15.3±9.9 (n=52) 13.9±8.0 (n=45)

p 0.001 0.001

Susceptible to CAZ Susceptible to FEP

20.5±14.8 (n=37) 20.6±14.4 (n=42)

14.1±7.7 (n=48) 15.5±9.2 (n=55)

0.011 0.035

In BSI group caused by cephalosporin susceptible strains, the

length of hospital stay was significant longer in group with CTX-M

positive than group with CTX-M negative with p < 0.05.

Septic shock

Mortality

45.0%

40.0%

p=0.482

p=0.406

p=0.380

p=0.786

35.0%

30.0%

25.0%

20.0%

5 / 2

1 1 / 4

15.0%

1 1 / 3

2 3 / 8

6 2 / 6

10.0%

2 3 / 7

5.0%

0.0%

3POS

NDM-1

3POS

NDM-1

Positive

36.4%

40.0%

27.3%

40.0%

Negative

23.1%

25.0%

23.1%

21.9%

Figure 3.6. Distribution of shock and mortality of patients caused by Carbapenem susceptible K. pneumoniae

In BSI patients caused by carbapenem susceptible K.

pneumoniae, the percentage of sock and mortality were higher in group

with 3-POS positive (36.4% and 27.3%) or NDM-1 positive (40% and

40%) than group with 3-POS negative (23.1% and 23.1%) or NDM-1

negative (25% and 21.9%), respectively.

Table 3.2. The length of hospital stays in BSI patients caused by carbapenem susceptible K. pneumoniae

Genotypic

Length of hospital stay (days) (Mean ± SD (n))

resistance

Positive

Negative

p

3-POS

38.2±21.7 (n=11)

16.4±11.2 (n=26)

0.001 NDM-1

44.8±23.5 (n=5)

19.5±14.5 (n=32)

0.002

In patients caused by carbapenem susceptible K. pneumoniae,

the length of hospital stay was longer in group with NDM-1 or 3-POS

positive than group with NDM-1 or 3-POS negative. The difference

was significance with p < 0.05.

DISCUSSION

General characteristics

The literature reviews present that BSI patients were reported

in elders with pre-existing conditions. The baseline of our participants

was similar to other reports in Vietnam and also in Asia with elder

patients related to pre-existing conditions. All most of patients were

reported primary source infections with E. coli-BSIs (83.5%) and K.

pneumoniae-BSIs (86.0%).

The typical subclinical of infections is increasing while blood

cells and other markers such as pro-calcitonin, serum lactate or

metabolic acidosis. The results from our study were similar and

relevant with other reports from Asia. This study was also assessed

SOFA score and mean of SOFA was over 3 points.

The ratio of shock and mortality of BSI patients depend on

different countries. The multiple-central reports presented 13% of

fatality of BSI patients and higher in intensive care units. We reported

proportion of shock and death were 16.5%-28% and 15.7%-24%,

respectively. We also evaluate medical burden of BSI via length of

hospital stay. That was 22.6 ± 17.7 days and similar to other reports

over the world.

The status of beta-lactam resistance and characteristics of

genes encoding beta-lactamases

4.2.1. The status of beta-lactam resistant Enterobacteriaceae

The observations of antibiotic resistant E. coli in Vietnam and

other countries of Asia present high prevalence of strains producing

ESBL and also resistance to cephalosporin and Quinolone. However,

they were still susceptible to carbapenem and aminoglycoside. Our

study observed E. coli producing ESBL was 58.3% and also high

resistance to AM (94.8%), CTX (70.4%) but still susceptible to TPZ

(80.9%). Other studies from Vietnam about ESBL-E. coli causing BSI

show that they were resistant to cephalosporin (45.0-45.6%);

carbapenem (0.8%), aminoglycoside (0.4-23.3%) and quinolone

(31.8-44.0%). The studies in Asia (including Vietnam) also perform

carbapenem resistant E. coli was about 0.3-1%, that was similar to our

result with 3/115 E. coli strains that was not susceptible to ETP.

The reports from Asia about K. pneumoniae causing BSI

demonstrate that the percentage of cephalosporin resistant K.

pneumoniae was not higher than E. coli. However, K. pneumoniae was

high resistant to carbapenem and still increasing. Our study showed

the similar results to other reports with 28-54% resistant to

cephalosporin and 26% resistant to carbapenem.

4.2.2. Characteristics of genes encoding beta-lactamases

The screening of 15 genes encoding beta-lactamase on E. coli

and K. pneumoniae, we observed over 90% of strains carrying at least

one of three genes encoding ESBL (CTX-M, TEM, SHV). The similar

to other reports in over the world, we observed on E. coli with highest

CTX-M (69.3%), follow-up TEM (58.3%) and SHV (0.9%). However,

SHV was detected in K. pneumoniae with highest prevalence (84%),

follow-up CTX-M (46% and TEM (46%). The reports from Asia show

that the most common subtype of CTX-M are CTX-M-14 and CTX-

M-15. Other meta-analysis of epidemiology of carbapenem resistant

mechanism over the world presents that KPC gene is the most common

in America and Europe but NDM-1 is the most common in India and

Southeast Asia. Our study reports the most common mechanism of

carbapenem resistance was NDM-1. The K. pneumoniae strains were

observed 38% of them carrying all of three genes (SHV, CTX-M and

TEM) and 18% of them carrying all of four genes (NDM-1, SHV,

CTX-M and TEM).

Correlation between genes encoding beta-lactamase and

phenotypic resistance and influence on the outcome

4.3.1. Correlation between genes encoding beta-lactamase and

phenotypic resistance

The most common mechanism of antibiotic resistance is

producing enzymes to hydrolyze antibiotics and the enzymes are

encoded by genes. Though the researches show that genes encoding

beta-lactamase are very variety with different function of subtype. The

mechanism of modulation of genotypic resistance is very complicated.

Consequently, it is not all of genes express to phenotypic resistance.

Our results of E. coli causing BSI, the distribution of CTX-M gene on

cephalosporin resistant strains was higher significance than

cephalosporin susceptible strains. The similar with K. pneumoniae, the

percentage of TEM and CTX-M on cephalosporin resistant strains was

higher significance than cephalosporin susceptible strains and the

distribution of NDM-1 and 3-POS was significant higher in

carbapenem resistant strains.

Our results show that the value of CTX-M on prediction of

cephalosporin resistance was 85.2%-100% sensitivity but 43.5%-

67.6% specificity. Other researches of genotypic and phenotypic

resistance perform high sensitivity and also specificity (over 95%).

However, the studies used whole genome sequencing method to

identify genotypic resistance, that was complicated and expensive

method to apply in clinical practice. The role of NDM-1 to predict

carbapenem resistant K. pneumoniae was high specificity (86.5%) but

moderate sensitivity (53.8%).

4.3.2. The influence of genes encoding beta-lactamase on the

outcome

The above results show that there were lots of strains carrying

CTX-M but still susceptible to cephalosporin. Analyzing correlation

between CTX-M and the clinical outcome in group caused by

cephalosporin susceptible strains show that the outcome was worse in

group with CTX-M positive than group with CTX-M negative (Septic shock, mortality was higher and the duration of fever after diagnosis, the length of hospital stay was significant longer). The results may be

explained by bacteria may increase expression of genotypic resistance

under the antibiotic pressure. So that when we choose antibiotic base

on phenotype, that may not suitable with genotypic resistance to make

worse outcome. Some researches based on whole genome sequencing

demonstrate that genetic resistance may be used to choose empirical

antibiotic therapy, especially severe infections such as sepsis.

Therefore, choice empirical antibiotic therapy base on genotypic

resistance like CTX-M and NDM-1 may get good value to improve the

clinical outcome in BSI patients caused by E. coli and K. pneumoniae.

CONCLUSION

Investigation of 165 BSI patients, including 115 patients

caused by E. coli and 50 patients caused by K. pneumoniae and

analyzing genotypic, phenotypic resistance and treatment response, we

could conclude the following:

1. Characteristics of beta-lactam resistance and distribution of

genes encoding beta-lactamase:

a. Charateristics of beta-lactam resistance

- The percentage of ESBL positive E. coli and K. pneumoniae were

58.3% and 22%, respectively; they were also resistant to Carbapenem

with 26.0% and 2.6%, respectively.

- K. pneumoniae and E. coli were high resistant to Cephalosporin but

still susceptible to penicillin plus beta-lactamase inhibitors.

b. Distribution of genes encoding beta-lactamases:

- Genes encoding ESBL: CTX-M, TEM and SHV were detected in E.

coli strains with proportion of 69.9%, 58.3% and 0.9%, respectively.

The similar in K. pneumoniae were 46.0%, 46.0% and 84.0%

respectively. The percentage of E. coli and K. pneumoniae strains

carrying at least one of three gene (SHV, CTX-M and TEM) were

91.3% and 90.0%, respectively.

- NDM-1 gene was detected in 24% of K. pneumoniae strains; 38.0%

of K. pneumoniae carrying all of three genes (SHV, CTX-M and TEM)

and 18% of strains carrying all of four genes (NDM-1, SHV, TEM and

CTX-M).

2. The value of genes encoding beta-lactamase to identify

phenotypic beta-lactam resistance of bacteria and influence of

genes encoding beta-lactamase on treatment response

a. CTX-M and NDM-1 were good candidates

to predict

Cephalosporin and Carbapenem resistance

- The sensitivity of CTX-M on prediction of cephalosporin resistance

to E. coli and K. pneumoniae was 84.9%-90.2% and 77.8%-89.5%,

respectively.

- The value of NDM-1 on prediction of carbapenem resistant K.

pneumoniae were 53.8% sensitivity and 86.5% specificity.

b. The influence of genes encoding beta-lactamase on treatment

response

- In BSI patients caused by cephalosporin susceptible strains, the

clinical outcome was worse in group with CTX-M positive (shock

21.6%-33.3% and mortality 18.9%-26.7%) than group with CTX-M

negative (shock 14.6%-19.2% and mortality 10.4%-15.4%), duration

of fever after diagnosis, length of hospital stay was significant longer

in group with CTX-M positive.

- In BSI patients caused by carbapenem susceptible K. pneumoniae:

the clinical outcome was worse in group with 3-POS positive or NDM-

1 positive (shock 36.4%-40,0% and mortality 27.3%-40.0%) than

group with 3-POS negative or NDM-1 negative (shock 23.1%-25.0%

and mortality 23.1%-21.9%). Longer length of hospital stay was in

group with NDM-1 positive or 3-POS positive.

PROPOSALS

Application of CTX-M and NDM-1 genes in combination

with antibiotic susceptibility test to guide empirical antibiotic therapy

for the treatment of serious infections caused by Eschierichia coli and

Klebsiella pneumoniae, such as bloodstream infections.

THE LIST OF PUBLICATIONS

RELATED TO THE THESIS

1. Trinh Van Son, Ngo Tat Trung, Nguyen Hong Nhung, Nguyen

Dang Manh và Le Huu Song (2020). The value of antibiotic

resistant genotype to predict cephalosporin resistance of

Klebsiella pneumoniae causing blood stream infections.

Journal of 108-Clinical Medicine and Pharmacy, (Volume 15) No3/2020: p141-148.

2. Trinh Van Son, Nguyen Dang Manh, Đao Thanh Quyen,

Nguyen Thi Kim Phương và Le Huu Song (2020). The value

of genotype in identifying carbapenem resistance of Klebsiella

pneumoniae causing blood stream infections. Journal of 108- Clinical Medicine and Pharmacy, (Volume 15) No4/2020:

p136-142.

3. Trinh Van Son, Ngo Tat Trung, Nguyen Dang Manh, Le Huu

Song (2020). Association of ESBL and CTX-M gene with the

clinical outcome of patients suffer from bloodstream

infections caused by Escherichia coli and Klebsiella

pneumoniae. Journal of 108-Clinical Medicine and pharmacy,

Vol.15 - Dec/2020: p. 49-53.