Occurrence and Genomic Characteristics of Hypervirulent Klebsiella pneumoniae in a Tertiary Care Hospital, Eastern India

Introduction

Klebsiella pneumoniae (KP), a major species within the K. pneumoniae complex of the genus Klebsiella and Order Enterobacterales, is a Gram-negative, encapsulated, non-motile bacterium, that is one of the most important and medically significant pathogen of humans causing a vast majority of infections in hospitals and communities throughout the world, including those of urinary tract, soft tissue, respiratory tract, and bacteremia.^1–3 Recently, the taxonomy of the K. pneumoniae complex has expanded with inclusion of several species and subspecies comprising of K. pneumoniae, K. quasipneumoniae subsp. quasipneumoniae, K. quasipneumoniae subsp. similipneumoniae, K. variicola subsp. variicola, K. variicola subsp. tropicalensis and K. africanensis.^4,5 Other members of the genus Klebsiella (apart from the K. pneumoniae complex) include K. aerogenes, K. oxytoca, K. grimontii, K. huaxiensis, K. michiganensis, K. pasteurii, K. spallanzanii and K. granulomatis.^2,4,5 Though this nomenclatural change has been attained with the aid of numerous molecular, genomic and proteomic tools available nowadays for better differentiation amongst the species within the K. pneumoniae complex, however, all taxa of this complex are still assigned as K. pneumoniae as the phenotypic and biochemical characteristics of all the members are extremely similar under standard microbiological conditions.^4,5 Classical K. pneumoniae (cKP) species ranks ninth among common agents of bloodstream infections and 2nd–4th among both hospital-acquired as well as community-acquired urinary tract infections (UTIs).^3,6 The organism was noted amongst the top ten pathogens (second only to Escherichia coli) in a cohort of community-onset bloodstream infections (BSIs) in a population-based surveillance study in rural provinces of Thailand over a period of 7 years from 2007–2014.⁷ In community acquired UTIs, it has accounted for 5–13.6% of isolates and in hospital-acquired UTIs, up to 20–58% of isolates.^6,8,9 In skin and soft-tissue infections, K. pneumoniae has accounted for 3.3%–8.1% cases.^10,11 Other less common infections include meningitis and enteritis in infants.^2,3

Hypervirulent Klebsiella pneumoniae (hvKP), a newly emergent modified KP strain, first evident as a significant pathogen in 1986 in Taiwan, differs from cKP strains in a lot of clinical and microbiological aspects.¹² Unlike cKP, which cause innumerable infections in hospitals as well as long-term health care facilities, hvKP are notorious as a cause of community-acquired infections in healthy and young individuals with an inclination for disseminated spread to faraway and distant sites.^12–15 The hvKP isolates produce considerable amounts of exopolysaccharide capsular material resulting in a hypermucoviscous phenotype (determined by a positive string test). In addition, most of these strains belong to capsular serotypes K1 and K2.^16–19 Other prevalent capsular serotypes have been noted to be K5, K20, K54 and K57.^20,21 Several putative virulence factors, such as mucoviscosity associated geneA (encoded by magA), regulator of mucoid phenotypeA (rmpA and rmpA2), adhesion type 3 fimbriae (mrkD), allantoin metabolism (allS) and various siderophores such as aerobactin (iuc), siderophore iutA (iutA), salmochelin (iroN), yersiniabactin (ybtS), and iron transport/phosphotransferase systems (kfuB) have been found to be associated with hvKP.^{17,18,22–24} Further, a notable difference is observed in their sensitivity profile to various antimicrobial agents, the hvKP strains being generally susceptible to most antibiotics in contrast to cKP which have been notorious for their increased propensity for acquiring antimicrobial resistance determinants. Especially, the occurrence of carbapenem-resistant K. pneumoniae strains characterized by acquired carbapenemase genes consisting of New Delhi metallo-ß-lactamase (NDM), oxacillinase (OXA), Klebsiella pneumoniae carbapenemase (KPC), imipenemase (IMP), Verona integron-encoded metallo-ß-lactamase (VIM) has made the therapy of cKP infections very taxing and extremely challenging.^25,26 In this scenario, the recent emergence of convergent hvKP strains in certain clinical settings with combined characteristics of high amounts of virulence and carbapenem resistance, is being viewed as an impending public health threat and an upcoming clinical crisis in recent times.^27–30

Increasing isolation of hvKP strains from other parts of the Asia Pacific Rim (apart from Taiwan) as well as from various regions of North and South America, Europe, Australia, and Africa, has truly made it a worldwide public health problem.^19,30–33 However, extremely limited information is available from India regarding the same.^31–34 Hence, we aimed to study the occurrence, antimicrobial susceptibility, and virulence characteristics of hvKP isolates from various clinical infections in a tertiary care hospital of eastern India. This information will help us in devising effective strategies for prevention, diagnosis, and treatment of hvKP infections as well as enhanced surveillance methods for rapid identification and monitoring of convergent K. pneumoniae strains.

Methods

This single-institutional cross-sectional study, approved by the Institutional Ethics Committee of All India Institute of Medical Sciences, Bhubaneswar, Odisha, India (Ref Number: IEC/AIIMS BBSR/PG Thesis/2019-20/05 dated 15th July 2019), and in compliance with the Declaration of Helsinki was conducted from August 2019 to June 2021 in a tertiary-care research, referral and teaching hospital in eastern India.

Isolate Identification and Susceptibility Testing

Consecutive non-repeat, discrete and clinically significant isolates of Klebsiella pneumoniae recovered from clinical specimens (blood, urine, pus, exudates, body fluid, respiratory samples, and any other significant source) of patients, receiving medical care, collected in a prospective manner and identified by VITEK-2 Compact automated platform (bioMérieux, Marcy-l’Étoile, France) were included in the study. Sample processing and interpretation were performed as per standard microbiological techniques, taking care to adhere to appropriate specimen collection procedures and other pre-analytical parameters.^35,36 Isolates from catheter tips and gastrointestinal tract were excluded to rule out colonizing isolates. For identification of hvKP, all confirmed K. pneumoniae isolates were subjected to the string test utilizing a metallic loop made sterile by flaming till it became red-hot and allowed to cool.^15,22 The loop was touched to the bacterial colony from a blood agar plate and stretched, with the test being considered positive upon generation of a viscous string >5 mm in length.^15,22

Antimicrobial susceptibility testing and interpretation was carried out as per the Clinical and Laboratory Standards Institute guidelines, primarily by the VITEK-2 Compact and supplemented by disc-diffusion/EzyMIC on Mueller-Hinton agar (HiMedia, Mumbai, India) wherever needed.³⁷ Colistin susceptibility testing was performed by the reference broth microdilution method.³⁷ Quality assurance of all the procedures, media, biochemical tests and antimicrobial susceptibility testing was maintained as per standard guidelines using control strains of Escherichia coli ATCC 25922, Pseudomonas aeruginosa ATCC 27853 and Staphylococcus aureus ATCC 25923 along with K. pneumoniae ATCC 700603 (SHV-18, Extended spectrum beta-lactamase positive), K. pneumoniae ATCC BAA-2146 (NDM-1, carbapenem-resistant) and K. pneumoniae ATCC BAA-1705 (bla_KPC positive).³⁷ Organisms showing intermediate resistance were included in the percentage of resistant organisms and those displaying non-susceptibility to at least one agent in three or more antimicrobial categories were defined as multi-drug resistant (MDR).³⁸

Molecular Investigations

Isolates testing positive by the string test (considered as hvKP) were subjected to genotypic characterization for detection of virulence-associated genes and capsular typing by polymerase chain reaction (PCR) using a panel of previously published oligonucleotide primer pairs (Sigma-Aldrich Ltd, St. Louis, Missouri) with their expected amplicon sizes as listed in Table 1.^{18–21,23,24} All carbapenem-resistant hvKP were further investigated for carbapenemase-encoding genes based on the four most prevalent carbapenemase types in India (NDM-1, OXA-48, OXA-181, KPC) by PCR (Table 1).²⁵ All the PCRs were performed as single-plex reactions.

Table 1 List of Primers for Virulence Gene Detection, Capsular Typing and Carbapenemase Gene Detection in Hypervirulent Klebsiella pneumoniae Strains

Briefly, genomic DNA was extracted by the HiPurA Bacterial Genomic DNA purification kit (HiMedia, Mumbai, India) as per manufacturer’s instructions and PCR was performed using DreamTaq Green PCR Master Mix (Thermo fisher Scientific, USA). Each 25 µL PCR reaction mixture was composed of 12.5 µL of the master mix, 1.25 µL of each forward and reverse primer solution (in a final concentration of 200 nM), 5 µL of template DNA with concentration of 100 ng/µL and 5 µL nuclease-free water to complete the final volume. Amplification was carried out in Veriti Dx 96-well Thermal Cycler (Thermo Fisher Scientific, USA) with cycling conditions consisting of initial denaturation at 94 °C for 5 mins, 35 cycles of denaturation (94 °C, 30 s), annealing (respective temperatures, 30 s) and extension (72 °C, 45 s), followed by a final extension for 8 mins at 72 °C. Amplified products of PCR were loaded onto 1% agarose gel stained with ethidium bromide and observed in an automated UV trans-illumination system (Syngene G: BOX, Synoptics, Cambridge, UK) for visualization of bands.

Data Extraction and Statistical Analysis

Patient-related information with relevant clinical and laboratory data were extracted from the medical records of patient case sheets, requisition slips, and admission charts. Microbiology test results were noted down as per the investigations conducted. Categorization as community- or healthcare-associated infection was based on the World Health Organization guidelines.³⁹ All data were entered in Microsoft Excel and analyzed using simple descriptive statistics consisting of ratios, proportions and frequencies. Chi-square analysis of contingency tables was used for categorical data and t test for continuous data wherever required. A p value < 0.05 was considered statistically significant.

Results

A total of 1004 K. pneumoniae were isolated during the study period with sample-wise distribution as follows: respiratory specimens consisting of sputum, endotracheal aspirates, and broncho-alveolar lavage (345, 34.3%), urine (326, 32.5%), pus and wound swabs (184,18.3%), blood (96, 9.6%), and body fluids consisting of cerebrospinal fluid, bile, peritoneal fluid, and pleural fluid (53, 5.3%). Eight hundred sixty-nine (86.5%) isolates were from patients admitted in the in-patient departments (IPDs) including intensive care units (ICUs), whereas 135 (13.4%) were from the outpatient departments (OPDs). Six hundred and twenty (61.7%) were from male patients, and 384 (38.2%) from females. The lowest and highest age at which K. pneumoniae was isolated was from the blood sample of a 2-day-old male child and from the urine sample of a 92-year-old male elderly patient, respectively.

Thirty-three isolates were identified to be hvKP with a resultant frequency of 3.3%. The demographic and clinical characteristics as well as the capsular serotypes, virulence-gene determinants and associated carbapenemase genes of these 33 hvKP isolates have been detailed in Table 2. Sample-wise occurrence of hvKP isolates was observed to be: respiratory specimens (15, 45.4%), followed by pus (8, 24.2%), urine (4, 12.1%), blood (3, 9.1%) and peritoneal fluid (3, 9.1%), while proportion-wise occurrence among K. pneumoniae isolates from various specimens were: body fluid, 5.6% (3/53); pus, 4.3% (8/184); respiratory specimens, 4.3% (15/345); blood, 3.1% (3/96), and urine, 1.2% (4/326). Eighteen (54.5%) hvKP were from patients admitted in various IPDs, eight (24.2%) from patients visiting the OPDs, and seven (21.2%) from patients admitted in the ICUs. Infection with hvKP was seen predominantly in male patients (25, 75.7%; male: female ratio, 3.1:1) and in adults in the age group of 18–60 years (22, 66.7%). Six (18.1%) hvKP infections were in elderly patients more than 60 years, and five (15.1%) among young patients less than eighteen years of age. One or more associated co-morbidities were noted in 22 (66.7%) patients, with underlying diabetes mellitus in 16 (45.7%) and a past history of pulmonary tuberculosis infection in eight (24.2%) (Table 2). Nineteen (57.6%) of the infections were healthcare-associated, while 14 (42.4%) were community-associated. Death was observed in 10 (30.3%) patients.

Table 2 Demographic Details and Genomic Characteristics of the 33 Hypervirulent Klebsiella pneumoniae Isolates, Eastern India

Analysis of capsular serotypes revealed that K2 was the predominant serotype detected in 11 (33.3%) isolates followed by K1 in 9 (27.3%) (Table 2). Other serotypes were, K54 in five (15.1%) and K5, K20, and K57 in two (6.1%) isolates each. Two (6.1%) isolates were non-typeable. As regards the virulence-encoding genes, adhesion type 3 fimbriae (mrkD) was the most prevalent virulence factor detected in thirty-one (93.9%), followed by siderophore (iutA) in twenty-eight (84.8%), and yersiniabactin (ybtS) in twenty-four (72.7%) isolates. Other virulence genes consisted of iron transport and phosphotransferase function (kfuB) in 22 (66.6%), aerobactin (iuc) in 21 (63.6%), regulator of mucoid phenotype A2 (rmpA2) in 19 (57.5%), rmpA in 16 (48.4%) and salmochelin (iroN) in 9 (27.2%) isolates (Table 2).

Overall, 216 (21.5%) KP isolates were susceptible to all the tested antimicrobial agents, whereas 788 (78.4%) were resistant to one or more agents. Six hundred forty three (64.0%) isolates were MDR. Table 3 shows the overall resistance of the isolates to the antibiotics tested as well as the comparative antibiogram analysis between the cKP and hvKP isolates, revealing significantly higher resistance in the former compared to latter for cephalosporins, amoxicillin-clavulanic acid, and fluoroquinolones (p < 0.05). Resistance to other antibiotics such as, aminoglycosides, co-trimoxazole, and carbapenems was also higher in the cKP isolates, though not statistically significant (Table 3). A total of 10 (30.3%) hvKP isolates were resistant to carbapenems indicating the occurrence of convergent hvKP strains with combined hypervirulence and carbapenem-resistance. Amongst the convergent hvKP strains, four (40%) were obtained from wound infections, two each (20%) from respiratory specimens and urine, and one each (10%) from blood and peritoneal fluid. All the ten carbapenem-resistant hvKP isolates were observed in hospitalized patients, with nine (90%) suggestive of hospital-acquired infection. Death rate was higher in patients harbouring an infection with carbapenem-resistant (convergent) hvKP strains compared to carbapenem-susceptible hvKP strains, though not statistically significant (4/10, 40% vs 6/23, 26.1%; p > 0.05). Genes encoding for carbapenemase-enzymes were detected in 6 (60%) of the ten carbapenem-resistant hvKP isolates distributed as bla_OXA-48 and bla_OXA-181 (3 isolates), bla_OXA-48 and bla_NDM-1 (1 isolate), bla_OXA-181 alone (1 isolate), and bla_OXA-48, bla_OXA-181 and bla_NDM-1 (1 isolate). Thus, overall oxacillinases (bla_OXA-48 and bla_OXA-181) were found in 5 (50.0%) and NDM (bla_NDM-1) in 2 (20.0%) isolates.

Table 3 Comparative Antimicrobial Susceptibility Profile of Classical and Hypervirulent Klebsiella pneumoniae Isolates to Various Antimicrobials

Discussion

This observational, cross-sectional study presents comprehensive data on K. pneumoniae infections, with special emphasis on hvKP isolates, in a tertiary care hospital in the eastern part of India, which receives patients not only from local areas on a primary basis but also large referrals from adjoining districts and states. During the study period, a total of 1004 consecutive, non-repeat, discrete KP isolates were obtained from various clinical samples received in our laboratory, with maximum recovery from respiratory tract specimens, followed by urine and pus. Previous studies from India, have shown maximum isolation of KP from urine samples, followed by sputum or exudative specimens.^31,40 Studies from China observed maximum isolation from blood followed by sputum and ascitic fluid¹⁷ or from pus followed by blood.¹⁸ Klebsiella pneumoniae in the current study were isolated mostly from hospitalized patients and less from OPD patients. Other studies in the past have also shown KP infections to be more common in hospitalized patients.^2,3,41 Overall frequency of hvKP in our study amongst the KP isolates was 3.3% similar to a previous study from south India reporting a low prevalence of 2.4%.³¹ In an Indian ICU, however, a higher rate of 19.4% (32/165) has been observed.⁴² Studies from southeast Asia report a moderate to very high prevalence of hvKP infections ranging from 12.3% to 42.4%.^{17,18,43–46} Other regions of the world such as, Canada, Iran, Mexico and France report a widely variable rate of 8.2%, 15.1%, 17.9% and 20.3%, respectively.^19,47–49 The difference in prevalence of hvKP in studies from across various continents may be due to the ethnicity of population, genetic susceptibility or local strain variation and needs to be studied in more detail. Demographic characteristics associated with hvKP infection in the current study such as male predominance and/or diabetes mellitus as an important underlying co-morbidity has also been ascertained in previous studies from India and outside.^17,18,31,42 Apart from diabetes, chronic diseases of the kidney, liver, and pancreas and underlying malignancies have also been found to be associated with hvKP infection.^{17,18,31,42,46} However, unlike our finding, past history of tuberculosis has not been looked into in any of the previous studies. Further, in congruence to our observation, the study from Indian ICU observed a majority (68.8%) of hvKP infections to be hospital-acquired, although mortality was much higher at 56.2% in that study.⁴²

The most frequent capsular serotype in the current study was K2 followed by K1, which is similar to various studies from China, South Korea and India with a predominance of K2 serotype.^{18,31,43,45,50} However, a study on whole genome analysis of seven hvKP isolates by Shankar et al in another study from India reported K1 capsular serotype in five isolates (71.4%) with no isolate belonging to K2 capsular serotype.³² Further, we found two hvKP isolates that were non-typeable. Such strains have also been reported from Iran and China where non-typeable capsular serotypes were found in 41% and 59.1% of their hvKP isolates, respectively.^17,47 Thus, there is a need to delineate newer capsular serotypes as these may be crucial in mapping hvKP infections. Information on capsular serotypes in our study will provide a baseline data in India which may be required in future to define the susceptibility status of hvKP isolates specifically in relation to specific capsular serotypes. A knowledge of capsular serotype will also help in designing vaccines in the potential scenarios of a future outbreak of antibiotic-resistant hvKP isolates.

Analysis of virulence-encoding genes revealed that mrkD was detected in thirty-one (93.9%) followed by iutA in twenty-eight (84.8%), ybtS in twenty-four (72.7%), and kfuB, and iuc in twenty-one (63.6%) hvKP isolates each. Similarly, a study from Vellore showed predominance of mrkD in all seven isolates which was followed by kfuB and alls in four (4/7, 57.1%).³² Mucoid phenotype regulator, rmpA2 was more frequently detected than rmpA in the current study. In a similar manner, rmpA2 was present in 8 isolates while 1 isolate harbored the rmpA gene in a study from north India.³³ Further, iron uptake encoding genes (fyuA/kfuA, kfuB) were predominantly observed in 7 isolates while that for allantoin metabolism (allB) was seen in a single isolate.³³ Siderophore-encoding genes (iroC, ybt, irp, iucA) were present in all the isolates³³ as observed in the current study. In China, ybtS (95.2%, 20/21), iutA (90.5%, 19/21) and iuc (57.1%, 12/21) were the most prevalent virulence factors,²⁴ while in Sudan, entB was the most predominant virulence gene (93.3%), followed by mrkD (78.3%), and kfu (60%).³⁰ In another very recently concluded study in China on 94 carbapenem-resistant KP, 89 (94.7%), 89 (94.7%), 87 (92.6%), and 13 (13.8%) isolates carried the wabG, fimH, uge, and kfu virulence genes, respectively, while, rmpA and magA were not detected.⁵¹ A finding of interest in the current study was that, virulence gene allS was found to be almost exclusively associated with the K1 serotype. This is in concordance with the results of Guo Y et al who state significantly higher positivity rates of ybtS and alls among K1 than in K2 isolates.¹⁸

The antimicrobial susceptibility profile of K. pneumoniae in the current study revealed high resistance rates to various antimicrobials (except colistin and tigecycline), comparable to published literature from different geographical locations of the world.^16,19 As expected, a higher resistance was observed amongst cKP compared to hvKP, statistically significant for cephalosporins, amoxicillin-clavulanic acid, and fluoroquinolones. Frequency of antimicrobial resistance in hvKP isolates was also high ranging from 25.0% to 48.4% (excluding colistin and tigecycline). Specifically, the frequency of convergent K. pneumoniae isolates was 1.0% (10/1004) amongst all K. pneumoniae strains or 30.3% amongst the hvKP (10/33) strains. These findings are extremely similar to a multicentric study in China with an overall prevalence of CR-hvKP of 1.1% (21/1838) among all K. pneumoniae isolates²⁴ as well as to the study by Bhardwaj et al⁴¹ and Chen et al⁵² with a carbapenem-resistance rate of 37.5% (12/32) and 23.8% (10/42) among hvKP strains respectively. However, the prevalent carbapenemase gene in the study by Zhan et al²⁴ was observed to be bla_KPC-2in contrast to bla_OXA-48-like, bla_OXA-181 and bla_NDM-1found in our study. A study from south India had reported both bla_NDM and bla_OXA-48 in two K2 serotype strains.³¹ Reports of CR- hvKP infections are also emerging from several other countries across the globe such as, United States of America, United Kingdom, Singapore, Russia, Iran and Japan.

Conclusion

The frequency of both hvKP and convergent hvKP infections in a tertiary care health-care setting in eastern India, appear to be low at present. However, continuous surveillance is necessary in order to detect any potential outbreak due to these entities in the future, as well as to study the outcomes in more detail so as to institute proper therapeutic strategies and appropriate infection control measures. Further, the present study will contribute to the existing but scarce data on the occurrence and genomic analysis of hypervirulent Klebsiella pneumoniae infections in India and would enable the clinical microbiology laboratory to strengthen its concept of hvKP and develop algorithms for routine screening and reporting of hypervirulent phenotypes. Information on capsular serotypes in our study will provide a baseline data which may be required in future to define the susceptibility status in relation to specific capsular serotypes and help in designing vaccines in the potential scenario of a future outbreak of antibiotic-resistant hvKP isolates.