Neonatal septicaemia caused by diverse clones of Klebsiella pneumoniae & Escherichia coli harbouring bla_CTX-M-15

Neonatal mortality in developing countries like India is very high12. Microbial infections leading to sepsis is a major contributor to neonatal deaths in the developing world3. The overall fatality rate due to neonatal sepsis and pneumonia in developing countries is estimated to be about 25 per cent, based largely on data for infants treated in hospitals3. Developing countries like India, Nigeria, Democratic Republic of the Congo, Pakistan, and China were responsible for large proportions of the global total for neonatal death due to sepsis4.

An important compounding factor in the treatment of these infections associated with sepsis is the emergence of extended-spectrum beta lactamase (ESBL)-producing Enterobacteriaceae5. ESBLs are derived from genes by mutation in the narrower-spectrum TEM-1, TEM-2 or SHV-1 β-lactamases67. Additionally, other types of ESBLs (CTX-M) that are plasmid-mediated, exhibit an overall preference for cefotaxime and ceftriaxone8. At present, more than 100 genetically distinct TEM, SHV, OXA and CTX-M-type ESBLs have been characterized (http://www.lahey.org/studies/webt.asp).

Among Gram-negative bacilli the two most important bacterial pathogens causing neonatal sepsis in developing countries are Klebsiella pneumoniae and Escherichia coli910. Studies on ESBL-mediated resistance mechanisms in adult patients from developing countries have been carried out1112131415161718. However, studies on the types of ESBLs and the clonal relationship of the organisms producing them, especially in neonatal infections from India are limited192021.

In this study, isolates of K. pneumoniae and E. coli causing bloodstream infection from neonates admitted to a tertiary care centre in New Delhi, India were analysed. The aim of the study was to carry out (i) molecular characterization of the ESBL genes in E. coli and K. pneumoniae isolates involved in neonatal sepsis, (ii) assessment of any significant differences in the clinical or demographical variables in the neonates that predispose them to acquisition of ESBL-producing organisms, (iii) determination of the molecular typing of the ESBL-producing isolates to ascertain the clonal carriage of the ESBL genes.

Material & Methods

Clinical setting and bacterial isolates: Isolates included in the study were collected from septicaemic neonates during January to May 2008 from Department of Microbiology at Vardhman Mahavir Medical College and Safdarjang Hospital, tertiary care centre in New Delhi. This centre has two nurseries, one for intramural admissions (born at the hospital) and one for extramural admissions (born outside the hospital). During the study period, a total of 1443 neonates were admitted to the nurseries, among whom, 405 neonates were suspected to have sepsis. Of these 405, 177 (43.7%) neonates were culture positive. Among these 177 neonates, 63 (35.59%) had aerobic Gram-negative bacilli (GNB) in their blood. Of the 63 GNBs, K. pneumoniae (n=16) and E. coli (n=10) were isolated. All K. pneumoniae and E. coli isolates obtained during this period were included in the study. Details of the neonates from whom K. pneumoniae and E. coli have been isolated are included in Table I.

Table I Demographic data and clinical features of the septicaemic neonates with E. coli or K. pneumoniae blood infection

Blood specimen collection and identification of the isolates: One ml blood for culture was drawn with aseptic precautions from a peripheral vein. Cultures were processed in BacT/ALERT 3D system (bioMeérieux, Marcy l’Etoile, France). For any culture which flagged positive, Gram stain was performed and subculture was done on appropriate medium based on the Gram stain: MacConkey agar and 5 per cent sheep blood agar for Gram-negative and Gram-positive organisms, respectively. Bottles were incubated in the system for up to seven days, at the end of which all negative bottles were subcultured once on blood agar before discarding. The isolates were identified by different biochemical tests and further verified by using an ID32E or ID32GN system (bioMeérieux).

Antibiotic susceptibility and MIC determination: Antibiotic susceptibility was determined as per Clinical and Laboratory Standards Institute (CLSI) criteria22. E. coli ATCC 25922 was used as quality control strain. The minimum inhibitory concentration (MIC) values (μg/ml) for cefotaxime, ceftazidime and cefepime were determined using E-tests (AB Biodisk, Solna, Sweden) and interpreted according to manufacturer's instructions.

ESBL confirmatory test: ESBL production was confirmed using the cephalosporin/clavulanic acid (cefotaxime/clavulanic acid and ceftazidime/clavulanic acid) combination disk test22. Fifty per cent of the isolates that were positive in the combination disk test were again confirmed by the double disk synergy test5.

PCR detection and sequencing of β-lactamases & integrons: Genotypic analysis for bla_SHV, bla_TEM, bla_OXA-1 and bla_CTX-M was performed for isolates found to be positive in ESBL confirmatory test and subsequently whole open reading frame (ORF) of PCR products were amplified for sequencing232425. PCR controls for bla_SHV and bla_TEM were K. pneumoniae ATCC 700603 and E. coli ATCC 35218, respectively. Control DNAs were used for detection of three groups of bla_CTX-M genes (provided by Dr Guillaume Arlet, Service de Bacteriology, Hospital Tenon, France). PCR was carried out to detect class 1, 2 and 3 integrons in all isolates and presence of plasmid-mediated ampC genes was also checked in cefoxitin-resistant isolates using PCR2627. PCR controls were used for ampC genes (MIR-1, DHA-1, ACT-1 and FOX-5) (kind gift of George A. Jacoby, Lahey clinic, Massachusetts). PCR products were purified using QIAquick PCR Purification Kit (Qiagen, USA), sequenced directly on both strands using the BigDye Terminator v^3.1 Cycle Sequencing Kit and analysed with an automated sequencer ABI3100 genetic analyser (Applied Biosystems, Foster City, CA). Sequence analysis was performed using the Lasergene DNASTAR sequence analysis software (DNAStar, Madison, USA). The BLASTN program was used for database searching (http://www.ncbi.nlm.nih.gov/BLAST/).

Pulsed-field gel electrophoresis (PFGE): PFGE was performed following the Pulse Net protocol (www.cdc.gov/pulsenet/protocols.htm). Overnight cultures of the isolates on Tryptic Soy Agar with 5 per cent sheep blood were suspended in the cell suspension buffer (100mM Tris, pH8.0, 100mM EDTA, pH 8.0). The optical density (600 nm) was adjusted to 1.3-1.4. Plugs were prepared by mixing 200 μl of the bacterial suspension [with 10 μl of Proteinase K (Takara, Japan)] with 200 μl of 1 per cent plug agarose (Seakem Gold Agarose, Cambrex, USA). Cell lysis was carried out in lysis buffer (50mM Tris: 50mM EDTA, pH 8.0, 1% Sarcosyl, 100 μg/ml Proteinase K) at 50°C. Digestion of DNA was performed by overnight incubation of the plug in the presence of 50 U/plug of the restriction endonuclease Xba I (New England Biolabs, USA) at 37°C. Electrophoresis was performed in a 1 per cent agarose (Pulse certified agarose, BioRad, Hercules, CA, USA) on a CHEF DRIII apparatus (Bio-Rad Laboratories) with 6V/cm for 19 h at 14°C with an initial switch time of 2.2 to 52.0 sec. Salmonella Braenderup H9812 was included as a molecular weight marker. XbaI macrorestriction patterns were compared visually and interpreted according to the criteria of Tenover et al28.

Conjugation: Conjugal transfer of bla_CTX-M to sodium azide-resistant E. coli J53 recipient (provided by George A. Jacoby, Lahey clinic, Burlington, Massachusetts, USA) was attempted by a broth mating assay29. Transconjugants were selected on Luria-Bertani agar plates containing cefotaxime (4 μg/ml) and sodium azide (100 μg/ml). The approximate sizes of the plasmids in the test isolates and transconjugants were estimated using logarithmic plots, generated with plasmids of known molecular mass30. E. coli K12 V517 (sizes of the plasmids are 54.2, 7.2, 5.6, 5.1, 2.7 & 2 kb) and Shigella flexneri YSH 6000 (sizes of the plasmids are 212, 3.9 & 2.7 kb) were used as reference standards. The presence of bla_CTX-M in the transconjugants was confirmed by PCR.

Statistical analysis: Fisher's Exact test was used for the assessment of risk factor for sepsis by comparing variables. This included sex, gestational age, birth weight, place of delivery (intramural/extramural), mode of delivery, use of a mechanical ventilator, type of feeding (feeding of expressed breast milk by enteral feeding tubes or breast feeding) and maternal risk factors (maternal pyrexia, meconium stained liquor, urinary tract infection, leaking per vaginum for more than 24 h duration, Hepatitis B positive, anaemia, preeclampsia and delivery of baby by traditional birth attendant). All comparisons were unpaired and all tests of significance were two-tailed.

Results

In comparison to K. pneumoniae, isolates of E. coli0 demonstrated higher resistance to the β-lactamase stable β-lactams, e.g. 3^rd generation cephalosporins (cefotaxime, ceftazidime, cefpodoxime, and ceftriaxone), monobactams (aztreonam), aminoglycoside (gentamicin), and fluoroquinolone (ciprofloxacin). On the other hand, K. pneumoniae isolates showed higher resistance than E. coli against ampicillin, piperacillin, amikacin and netilmicin. Only imipenem exhibited 100 per cent activity in both isolates. Rates of resistance to β-lactam and non-β-lactam antibiotics were as follows (in K. pneumoniae versus in E. coli, respectively): ampicillin, 100 vs 88 per cent; piperacillin, 100 vs 88 per cent; cefotaxime, 60 vs 75 per cent; ceftazidime, 30 vs 50 per cent; cefpodoxime, 60 vs 75 per cent; ceftriaxone, 60 vs 75 per cent; azteonam, 60 vs 75 per cent; amikacin, 40 vs 25 per cent; gentamicin, 60 per cent vs 75 per cent; netilmicin, 50 vs 19 per cent; and ciprofloxacin, 60 vs 91 per cent. Cefoxitin resistance was low in both K. pneumoniae [30% (3/10)] and E. coli [6% (1/16)] isolates. All ESBL-negative isolates showed MIC values for cefotaxime, ceftazidime and cefepime in the range of 0.094 - 0.5 μg/ml and values for ESBL-positive isolates are presented in Table II.

Table II Determination of MIC values and sequence analysis of ESBL genes in ESBL-positive K. pneumoniae and E. coli isolates

The overall percentage of ESBL-positive isolates in cephalosporin/clavulanic acid disk test was 60 per cent (6/10) for K. pneumoniae and 75 per cent (12/16) for E. coli. Isolates of K. pneumoniae and E. coli which were positive for ESBL by phenotypic tests showed presence of bla_CTX-M group 1. Among ESBL-positive E. coli (n=12), bla_OXA-1 was present in 100 per cent of isolates but none showed the presence of bla_SHV and bla_TEM. ESBL-positive K. pneumoniae (n=6) isolates possessed bla_OXA-1[83% (5/6)], bla_TEM [100% (6/6)] and bla_SHV [67% (4/6)]. Sequencing revealed the presence of bla_TEM-1, bla_SHV-1, bla_SHV-28, bla_SHV-11, bla_SHV-12 and bla_CTX-M-15 in ESBL-positive isolates (Table II). Plasmid- mediated ampC β-lactamases could not be detected in any of the cefoxitin-resistant isolates (n=4).

Only seven ESBL-positive isolates (E5, E31, E55, E78, E98, E104 & E113) were found to possess class 1 integron which was confirmed by sequencing. However other classes of integrons could not be identified in any of these strains. Molecular typing of K. pneumoniae (n=10) revealed diversity (Fig. 1). Among E. coli (n=16) isolates, one clonal cluster was identified which consisted of three isolates (E5, E10 & E31). Another cluster of E. coli consisted of two (E27 & E34) closely related isolates (Fig. 2). CTX-M-15 was detected in all isolates that were ESBL-positive irrespective of their genotypes.

Close

Fig. 1

Pulsed-field gel electrophoresis (PFGE) of XbaI-digested genomic DNA of CTX-M-15 producing K. pneumoniae (Lane 1-10) isolated from blood of neonates. Lane M: Salmonella serotype Braenderup H9812 as reference standard (band sizes in kilobases).

Export to PPT

Close

Fig. 2

Pulsed-field gel electrophoresis (PFGE) of XbaI-digested genomic DNA of CTX-M-15 producing E. coli (Lane 1-16) isolated from blood of neonates. Lane M: Salmonella serotype Braenderup H9812 as reference standard (band sizes in kilobases).

Export to PPT

Different sizes of megaplasmids (~54.2 Kb upto >212 Kb, data not shown) were isolated from ESBL-positive isolates (12 E. coli & 6 K. pneumoniae) hourbouring bla_CTX-M-15 gene. Conjugal transfer of bla_CTX-M-15 gene carried out for one isolate each of E. coli and K. pneumoniae was successful. The presence of bla_CTX-M-15 was confirmed by PCR in the transconjugants and plasmid analysis indicated that these had acquired megaplasmids present in the donor isolates (Fig. 3).

Close

Fig. 3

Plasmid profiles of test isolates and transconjugants of E. coli and K. pneumoniae hourbouring blaCTX-M-15. Lanes 1 & 2: transconjugant of E. coli and test isolate of E. coli from blood of neonates, respectively. Lanes 3 & 4: transconjugant of K. pneumoniae and test isolate of K. pneumoniae from blood of neonates, respectively. Lanes 5 & 6: E. coli K12 V517 and Shigella flexneri YSH 6000 were used as reference standards respectively.

Export to PPT

Risk factors for sepsis due to ESBL-producing organisms were examined for neonates (Table III). Among all the factors analysed for acquisition of ESBL-producing organisms, use of enteral feeding tubes was found to be the only significant risk factor (P=0.02). No significant association was found between ESBL production in this study group and mortality of neonates.

Table III Analysis of risk factors associated with sepsis due to ESBL-producing organisms in neonates (n=23)

Discussion

There is limited information about the diversity of the ESBL-producing genes and the clonality of the organisms involved in neonatal infections in India. It is necessary to know whether the same bacterial clones are responsible for the infections or whether dissemination of a particular ESBL-gene via genetically unrelated bacterial clones has been taking place. This study carried out in neonates highlights the presence of ESBL-producing K. pneumoniae and E. coli of diverse clonality in neonatal infections. The rate of resistance to most antibiotics was alarming, suggesting that the WHO recommended ampicillin and gentamicin combination as first line treatment of neonatal sepsis may no longer be effective31. Resistance to 3^rd generation cephalosporins and monobactam was seen in these isolates, being higher amongst E. coli isolates. Ciprofloxacin resistance was more frequent in E. coli isolates, as reported previously32. Carbapenems which are reserve drugs for treatment of neonatal sepsis, showed 100 per cent sensitivity.

bla_SHV genes were identified only in K. pneumoniae isolates. All different bla_SHV genes identified were non-ESBL except for SHV-12 which is an ESBL type. In addition, TEM-1 was found only in K. pneumoniae and the occurrence of bla_TEM-1 genes were more than bla_SHV genes in K. pneumoniae isolates which is similar to a previous study on neonatal sepsis20. The E. coli isolates thus harboured a different set of β-lactamase genes in comparison to the K. pneumoniae isolates.

The majority of the bla_TEM and bla_SHV genes present were the classical types and were not responsible for the ESBL phenotype. It was the bla_CTX-M-15 gene that was responsible for the ESBL phenotype. PFGE of the CTX-M-15-producing isolates revealed vast diversity and no epidemic clone could be seen. The clonal diversity and the possibility of dissemination of such plasmid mediated genes raise concern as experiments in the laboratory show that transfer of such genes can occur by conjugation. In addition, the presence of integron-1 in some isolates probably allows dissemination of bla_CTX-M-15 gene.

Three K. pneumoniae and one E. coli isolates were found to be cefoxitin-resistant though plasmid-mediated ampC β-lactamases were not detected. This could probably occur due to loss of porins in the isolates or the presence of chromosomally-encoded ampC β-lactamases in the case of E. coli33.

Infection with ESBL-producing organisms was significantly more among the neonates for whom expressed breast milk was fed through enteral feeding tubes in comparison to the neonates fed at the breast. Previous study showed that the use of the enteral feeding tube was an independent risk factor for colonization by multiple antibiotic-resistant K. pneumoniae and E. coli but the subjects were not neonates34. Further studies with a larger population need to be done to establish this association.

Though the study is limited by the small sample size, it highlights the presence of CTX-M-15 in clonally diverse K. pneumoniae and E. coli isolates indicating that bla_CTX-M-15 is probably disseminated horizontally. The high prevalence of ESBL organisms and a transmissible resistance gene (bla_CTX-M) is of great concern in a country with high population density and infant mortality rate. Though carbapenems remain the most active antibiotics, the high rate of ESBL-producing organisms may compel clinicians to increase the usage of carbapenems, with a probable increase in carbapenem resistance. The spread of ESBL-producing bacteria necessitates a change towards ‘evidence-based’ treatment practices, including prudent use of antibiotics.