Detection of hypervirulent Klebsiella pneumoniae (hvKp) strains directly from spiked blood cultures using a commercial Loop-Mediated isothermal amplification (LAMP) assay

Abstract

Background

Hypervirulent K. pneumoniae (hvKp) strains are characterized by their enhanced ability to evade immune responses and disseminate systemically. Rapid identification of hvKp strains is critical for guiding clinical management and implementing effective infection control measures. Loop-Mediated Isothermal Amplification (LAMP) assays provide a rapid and cost-effective method for detecting bacterial pathogens. This study evaluates the performance of the Eazyplex^® hvKp assay for the direct detection of hvKp strains from spiked blood cultures.

Methods

We collected 20 K. pneumoniae (Kp) isolates between December 2021 and August 2024 from two hospitals in Northern Spain. Capsular serotyping and virulence gene detection were performed using PCR and whole-genome sequencing (WGS). The Eazyplex^® hvKp LAMP assay was tested on spiked blood cultures inoculated with hvKp isolates. Virulence profiles were assessed using the Kleborate scoring system.

Results

Seventeen Kp isolates had a Kleborate score of ≥ 2, suggesting high virulence. The LAMP assay detected 87 out of 95 virulence targets, demonstrating an overall accuracy of 91.5%. Although eight target genes were not directly detected, fluorescence signals indicated amplification in all cases. The assay identified 16 strains with high virulence profiles (score ≥ 3), with 14 strains scoring 4 or 5. The LAMP-based test effectively detected hvKp directly from blood cultures, with time-to-results ranging from 6:43 to 17:11 min.

Conclussion

The Eazyplex^R hvKp LAMP assay is a rapid and effective method for identifying hvKp strains directly from blood cultures. This study supports its potential utility in clinical microbiology for early detection and epidemiological surveillance of hvKp infections. However, limitations in the Kleborate scoring system indicate that additional virulence biomarkers may be needed to improve the accuracy of hvKp classification.

Introduction

Klebsiella pneumoniae (Kp) is a clinically relevant gram-negative bacterium that causes opportunistic infections such as urinary tract infections, liver abscesses or bacteremia, among others [1]. Two different pathotypes, the carbapenem-resistant Kp (Cr-Kp) and the hypervirulent-Kp (hvKp) have evolved from the classical-Kp (cKp) as a major cause of hospital and community-acquired infections, respectively [2]. There is currently no consensus on the precise definition of the hvKp pathotype. However, it is widely accepted that hypervirulent strains exhibit increased siderophore production, capsule overexpression, and a tendency to cause invasive infections. In response to the diagnostic challenge of differentiating hvKp from cKp, the Kleborate scoring system was proposed, which evaluates the presence of siderophores such as aerobactin, yersiniabactin, and colibactin [3]. Aerobactin (iuc) is the strongest predictor of the hypervirulent (hvKp) pathotype and, when present, confers a virulence score of ≥ 3, the threshold typically used to define hvKp. In contrast, yersiniabactin (ybt) and colibactin (clb) can be found in both classical and hypervirulent strains [4], contributing lower virulence scores of 1 and 2, respectively. While these loci are associated with virulence, they alone are not consistent indicators of hypervirulence, unlike aerobactin, which is the major determinant in the Kleborate score. However, this approach does not consider other virulence-associated markers such as rmpA or rmpA2, which are frequently found in hvKp [5]. Since both the hvKp pathotype and bloodstream infections are major causes of morbidity and mortality [6], rapid identification of these strains could be crucial in the future when targeted treatments against this pathotype become available.

Several methods have been implemented for identification of hvKp strain such as conventional PCR, the string test, or the Galleria mellonella larvae infection model [7, 8]. Loop-mediated Isothermal Amplification (LAMP) is a promising tool that amplifies DNA with high efficiency and rapidity under isothermal conditions. Additionally, it is a cost-effective tool, making it suitable for inclusion in diagnostic workflows or even for point-of-care testing. Indeed, this method has been previously proven for the rapid identification of pathogens and antimicrobial resistance genes [9, 10]. The “EazyplexR hv-K. pneumoniae” (Amplexdiagnostics GmbH, Gars, Germany) is a commercially available RUO assay that detects six virulence genes (iucC, iroC, rmpA, rmpA2, ybt and clb) from colonies and provides the Kleborate score [3]. It is important to note that the assay is validated for use with cultured colonies only and is not recommended for direct testing of clinical specimens. In this study, we extended its application by comparing the LAMP assay results with those obtained through conventional PCR and whole-genome sequencing (WGS). The isolates used in this study had been previously identified and genetically characterized by PCR and WGS as part of routine surveillance and research initiatives conducted in our laboratory. The objective of this study was to evaluate the performance of the Eazyplex^R “hv-K. pneumoniae” kit (Amplexdiagnostics GmbH, Gars, Germany) for the direct detection of hvKp strains from spiked blood cultures, simulating clinical bacteremia. The isolates selected had been characterized in our laboratory to establish well-defined virulence profiles for assay validation.

Methods

Collection, identification and antimicrobial susceptibility testing (AST) of the Kp isolates

We collected 20 Kp isolates between December 2021 and August 2024 from two different hospitals in Northern Spain. Among the 20 strains analysed in this study, two Cr-Kp strains presenting only ybt were included as negative controls (strains VHVKP12 and VHVKP13, typed as K25-ST17 and K102-ST307, respectively). The isolates were obtained from different clinical specimens, including bronchial aspirate (n = 1), liver abscesses (n = 4), other abscess locations (n = 1), bile (n = 1), blood cultures (n = 4), urine (n = 4), surgical wound (n = 1), sputum (n = 3), and ear drainage (n = 1) (Table 1). They were identified using the Vitek MS matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry system (BioMérieux, Marcy-l’Étoile, France). Antibiotic susceptibility testing was performed using AST-N426 cards for Enterobacterales in the VITEK^R 2 system (BioMérieux, Marcy-l’Étoile, France), and the results were interpreted according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines [11]. The string test was conducted to assess the mucoviscosity of the strains [12].

Table 1.

Summary of the LAMP-assay results, K-TYPE, MLST and source of the strains

Strain	K-TYPE y MLST	iucC	iroC	rmpA	rmpA2	ybt	clb	Conventional PCR	Source	Kleborate score
BUH1	K2-ST66	12:11	17:11	10:22	10:00	7:43	8:48	iucC, iroC, rmpA, rmpA2, ybt, clb	Bronchial aspirate	5
BUH2	K1-ST23	10:46	10:10	10:04	9:08	- (curve)	7:34	iucC, iroC, rmpA, rmpA2, ybt, clb	Liver abscess	5
BUH3	K1-ST23	12:57	11.35	- (curve)	10:38	- (curve)	8:33	iucC, iroC, rmpA, rmpA2, ybt, clb	Liver abscess	5
BUH4	K1-ST23	11:53	12:16	8:53	9:52	7:18	6:58	iucC, iroC, rmpA, rmpA2, ybt, clb	Liver abscess	5
BUH6	K57-ST412	12:03	11:07	11:05	8:54	–	–	iucC, iroC, rmpA y rmpA2	Bile	3
K30B1	K30-ST198	–	12:45	- (curve)	–	7:48	–	rmpA, ybta, iroC	Liver abscess	1
K2-HUMV1	K2-ST66	9:45	11:46	8:22	–	7:03	7:58	iucC, iroC, rmpA, rmpA2, ybt, clb	Blood culture	5
VHVKP1	K1-ST23	11:39)	11:06	11:08	10:33	9:11	7:04	iucC, iroC, rmpA, rmpA2, ybt, clb	Urine	5
VHVKP2	K1-ST23	- (curve)	11:57	10:23	10:10	7:42	6:50	iucC, iroC, rmpA, rmpA2, ybt, clb	Abdominal abscess	5
VHVKP3	K2-ST6997	10:12	9:54	7:24	–	6:19	–	iucC, iroC, rmpA, ybt	Surgical wound	4
VHVKP4	K1-ST23	8:24	8:23	8:51	8:16	6:22	6:43	iucC, iroC, rmpA, rmpA2, ybt, clb	Sputum	5
VHVKP5*	K54-ST714	11:57	12:38	- (curve)	8:59	9:02	–	iucC, iroC, rmpA, rmpA2, ybt	Urine	4
VHVKP6	K2-ST380	- (curve)	10:31	9:30	–	9:07	–	iucC, iroC, rmpA, ybt	Blood culture	4
VHVKP7	K2-ST380	11:05	10:00	10:36	–	8:53	–	iucC, iroC, rmpA, ybt	Sputum	4
VHVKP8	K1-ST23	–	11:50	–	–	10:53	8:51	iroC, ybt, clb	Urine	2
VHVKP9	K2-ST65	- (curve)	13:18	13:05	10:10	–	–	iucC, iroC, rmpA, rmpA2	Blood culture	3
VHVKP10	K62-ST45	7:56	-	-	-	7:49	–	iucC, ybt	Blood culture	4
VHVKP11*	K54-ST714	12:13	11:14	9:00	9:52	8:57	–	iucC, iroC, rmpA, rmpA2, ybt	Sputum	4
VHVKP12 (NC)	K25-ST17	–	–	–	–	9:59	–	ybt	Ear	1
VHVKP13 (NC)	K102-ST307	–	–	–	–	8:35	-	ybt	Urine	1

Additionally, we included the virulence scores of the study strains and the virulence genes proposed by the Kleborate [8], which were detected by conventional PCR. The positive results of the LAMP-based assay are reported as the time to detection, while negative results are indicated as a hyphen (-). Strains VHVKP12 and VHVKP13 were used as negative controls. Although the kit failed to detect eight target genes, a fluorescence curve was observed in all cases. These cases are shown as “- (curve)”. Strains marked with an asterisk ((*) caused two distinct episodes with different clinical presentations, occurring one and a half months apart

Detection of capsular serotypes and virulence genes by conventional PCR

After identification of the hypermucoviscous (hmv) strains, we performed a second screening using conventional PCR to detect hvKp strains. This included various capsular serotypes (K1, K2, K5, K20, K54 and K57) and virulence-associated genes (rmpA, iroC, alls, kfuBC, wcaG, iucB, iroNB, ureA, wabG, uge, fim and ybtA), as proposed by Jian-Li W et al. [13]. The PCR products were analysed on a 1.5% agarose gel electrophoresis, and the results were compared with those of positive control strains harboring the targeted virulence genes.

Whole-genome sequencing (WGS) analysis

The DNA of the isolates was extracted using the Qiagen DNeasy^® Blood and Tissue Kit (Qiagen, Hilden, Germany) and sequenced on a MinION-Mk1B device (Nanopore Technologies, Oxford, UK) for 24 h using the Rapid Barcoding Kit (SQK-RBK004). The isolates were typed with the Kaptive 2.0 [14] and BIGSdb [15] databases. Virulence genes were identified by using BLASTn by using the rmpA, rmpA2, iuc, iro and peg-344 genes obtained from NCBI^R.

Eazyplex^R HvKp LAMP assay and kleborate score

All isolates used for the assay were processed identically by inoculating them into artificially spiked blood cultures for assay evaluation. No comparisons were made between strains based on their original site of isolation, as the study was not designed to assess performance variability related to specimen origin.

We inoculated 200 µl of a 0.5 McFarland bacterial suspension of hvKp isolates into BacT/ALERT^R FA Plus bottles (BioMérieux, Marcy-l’Étoile, France) with 5–10 ml of healthy donor blood, following the protocol described by Vidal-García M et al. [16]. The spiked blood cultures (BC) were incubated in the BacT/Alert^R Microbial Detection System (BioMérieux, Marcy-l’Étoile, France) until they tested positive. Then, we suspended 25 µl of positive samples into 500 µl of resuspension and lysis fluid (RALF) buffer solution (supplied with the kit) and boiled them for 2 min for thermal lysis, following the instructions of the Eazyplex^R SuperBug CRE kit (Amplexdiagnostics GmbH, Gars, Germany). This kit is a molecular rapid test designed to detect carbapenemase-producing bacteria and has been validated for BC samples.

Finally, we centrifuged the suspension for 30 s at 1000 x g, and 25 µL of the supernatant was added to each tube of the eazyplex test strip. Tests were run on a Genie HT machine (Amplexdiagnostics GmbH, Gars, Germany) at 65 °C for 20 min. The virulence of the Kp isolates was evaluated using the Kleborate score (iuc, ybt and clb) rmpA, rmpA2 and iroC were also detected by the “Eazyplex^R hv-K. pneumoniae” kit (Amplexdiagnostics GmbH, Gars, Germany), however, these genes do not contribute to the Kleborate score.

The hvKp isolates used in this study were obtained from various clinical specimen types, including bronchial aspirates, liver abscesses, blood cultures, and urine. However, all isolates were processed identically by inoculating them into artificially spiked blood cultures for assay evaluation. No comparisons were made between strains based on their original site of isolation, as the study was not designed to assess performance variability related to specimen origin.

Results

Seventeen of the 20 Klebsiella pneumoniae (Kp) isolates analyzed in this study had a Kleborate virulence score of ≥ 2, indicating increased virulence [3]. The results of the “Eazyplex^R hv-K. pneumoniae” LAMP assay are presented in Table 1, including the elapsed time required to obtain results, calculated as the threshold period necessary for the appearance of fluorescence. The time to results ranged from 6:43 to 17:11 min for the different target genes. The kit accurately detected 80 out of 88 target genes. Although this LAMP-based assay failed to show a positive result for eight target genes, a fluorescence curve was observed in all cases, and the fluorescence intensity threshold was reached in every instance. The assay showed that 16 out of 20 strains had a high virulence profile with scores ≥ 3. Additionally, 14 out of these 16 strains scored 4 or 5. The two remaining strains, VHVKP8 and K30B1, were also included in this study due to their atypical virulence profiles and clinical relevance, despite not meeting the standard Kleborate criteria for hvKp. The VHVKP8 belonged to the K1-ST23 sublineage, which is classicaly associated with hypervirulent strains. Notably, this strain carried the same IncFIB(K)/IncHI1B virulence plasmid replicon but lacked the region containing the rmpA, rmpA2, peg-344 and iuc and iro genes. The VHVKP8 strain only carried ybt and clb as virulence genes according to the LAMP-based assay, however, we also detected other virulence genes such as magA, mrkD, ureA, fim, allS, uge, wabG and kfu through conventional PCR and WGS (Table 2). Finally, despite having a virulence score of 1, the K30B1 strain was also included in this study as it harboured ybt, iroBCDN and rmpA genes. Additionally, this strain caused three episodes of liver abscesses in the same patient over a period of 9 months, which is a typical clinical presentation of the hvKp pathotype. It is worth noting that the VHVKP5 and VHVKP11 strains caused two distinct episodes with different clinical presentations in the same patient (urinary tract infection and pneumonia), occurring one and a half months appart.

Table 2.

Virulome of the different strains

In addition to rmpA, rmpA2, iroC, iucC, ybt and clb, other virulence-associated genes such as peg-344 or magA [17], were also detected in the study strains by conventional PCR and WGS. While magA was exclusively found in the seven K1-ST23 strains, peg-344 was detected in 15 out of 17 strains. The complete virulence profile of the strains, detected by combining conventional PCR and WGS results, is presented in Table 2. Regarding AST, all strains except BUH6 and VHVKP8 were susceptible to most antibiotics except ampicillin, due to the presence of an intrinsic bla_SHV gene in all strains. However, BUH1 and K2-HUMV1 strains lacked the bla_SHV gene and were susceptible to ampicillin. The VHVKP8 strain carried bla_CTX-M-15, bla_OXA-1, bla_TEM-1, dfrA14 and qnrB1 genes, conferring resistant to second, third and fouth generation of cephalosporins, amoxicillin/clavulanic acid, piperacillin/tazobactam, quinolones and trimethoprim-sulfamethoxazole, respectively. This strain was susceptible only to carbapenems, aminoglycosides, and nitrofurantoin. The BUH6 strain also carried antimicrobial resistance genes, including qnrs1, tet(A) and bla_LAP-2, being also resistant to amoxicillin/clavulanic acid, quinolones and tetracyclines.

Discussion

Since the emergence of hvKp strains in 1980 in Asian countries, this pathotype has spread to western countries [5], becoming a global health problem. A recent report by the European Centre for Disease Prevention and Control (ECDC, https://www.ecdc.europa.eu/en) highlighted an increase of infections caused by K1-ST23 hvKp strains carrying carbapenemase genes across 10 European countries [18]. Thus, it is crucial to develop rapid assays to improve surveillance and the early detection of hvKp strains. Due to its rapid and targeted nature, conventional PCR was used as a preliminary screening tool to detect selected virulence genes commonly associated with hvKp strains. In contrast, WGS enabled a more comprehensive analysis, allowing for in-depth characterization of the isolates, including the identification of the genetic context of virulence factors (e.g., plasmid or chromosomal location), and the presence of additional or uncommon markers. The combined use of PCR and WGS increased the accuracy and robustness of strain characterization. LAMP-based assays could be up to 10 times more accurate than conventional PCR for detecting target sequences in selected cases [10]. Indeed, Liao W et al. assessed the use of LAMP by detecting the peg-344 gene in 28 strains, demonstrating 100 times greater sensitivity than conventional PCR. In addition, Rödel J et al. evaluated the combined use of Eazyplex^R hvKp and SuperBug CRE assays to detect both hvKp strains and hybrid clones, following the risk assessment from the ECDC [19]. They identified 14 strains with a virulence score ≥ 2.

At present, there is no universally accepted definition of the hypervirulent pathotype. However, most authors concur about its link to the enhanced siderophore production, the hyper-expression of the capsule and the invasive infections [5]. The Kleborate score, as proposed by Lam et al. [3]., is an attempt to normalise the identification of hvKp based on siderophore genes (iuc, ybt and clb), but it does not consider rmpA or rmpA2, as both are commonly truncated. However, ybt and clb are also detected in classical and carbapenem-resistant isolates [20], which limits interpretation. For instance, strain VHVKP8 had a score of 2 (ybt and clb only), but failed to have the core hvKp features. In contrast, strain K30B1 was scored 1 but was positive for rmpA, iroBCDN, and other virulence genes and caused recurrent liver abscesses [21], which suggests that the current scoring may under-estimate clinically relevant hypervirulence. Russo et al. have raised doubts on the sensitivity of the Kleborate score compared with a biomarker-based method using iucA, iroB, peg-344, rmpA and rmpA2 [22]. Among our isolates, 12 harbored all five biomarkers and 2 strains missed only rmpA2, corresponding to a better fit with the hvKp profile. These observations highlight the importance of choosing appropriate detection strategies when testing hvKp in clinical settings.

In this context, the Eazyplex^R “hv. K. pneumoniae” kit (Amplexdiagnostics GmbH, Gars, Germany) may offer a practical and rapid alternative to conventional PCR. While its performance in spiked blood culture is promising, further prospective evaluation using clinical samples is necessary to fully assess its diagnostic value and potential role in surveillance. Although currently validated only for use with cultured colonies, our results suggest that this kit may provide rapid answers in microbiology laboratories. Its direct application to positive blood cultures could enable early identification of hvKp strains, facilitating timely clinical decision-making even in the absence of complete phenotypic or genotypic characterization. In the clinical setting, the assay could be applied in the microbiology laboratory as a point-of-care or incorporated into standard blood culture workflows to support early diagnosis of hvKp, and facilitate infection control practices. Finally, while the use of spiked blood cultures provides a controlled setting for performance assessment, it does not fully replicate the complexity of clinical samples, where lower bacterial loads or co-infections may interfere with detection. Therefore, further clinical validation is essential to confirm the assay’s diagnostic robustness.

Conclusion

This study demonstrates that the Eazyplex^R hvKp LAMP assay kit can rapidly identify hvKp strains directly from spiked blood cultures. This capability is clinically significant, as bacteraemia caused by hvKp strains is associated with high morbidity and mortality [4]. However, while we have shown that this assay accurately detect virulence genes in K. pneumoniae, some limitations remain. Relying on the Kleborate score alone may misclassify some strains as hvKp, particularly due to ybt and clb detection. Adjusting the assay to focus on peg-344 instead may improve hvKp identification accuracy. Although the assay showed good performance in detecting hvKp in spiked blood cultures, further prospective studies using real clinical samples are needed to confirm its usefulness in routine practice. These studies should evaluate how well the assay works in more complex situations, such as polymicrobial infections or samples with low bacterial loads, which better reflect real clinical scenarios.

Representation of the virulence genes in hvKp isolates detected by conventional PCR and WGS analysis. Blue: presence of a virulence gene; white: absence. The underlined genes are those detected by the “Eazyplex^R hvKp” LAMP assay. Strains marked with an asterisk (*) were isolated from the same patient during different episodes. NC: negative controls.

Author contributions

- Domingo Fernández Vecilla and Alain Ocampo Sosa: Writing – Original Draft Preparation; Methodology; Conceptualization. - Nuria Fraile Valcárcel and Jorge Rodríguez Grande: Data curation; Formal Analysis. - María Catalina Lomoro: Methodology. - Zaira Moure García, Sergio García Fernández, María Siller Ruiz, María Pía Roiz Mesones, José Luis Díaz de Tuesta del Arco, Mikel Joseba Urrutikoetxea Gutiérrez, María Carmen Fariñas: Writing – Review & Editing. - Domingo Fernández Vecilla, Jorge Rodríguez Grande and Nuria Fraile Valcárcel, equal contribution. - All the authors have read and approved the manuscript by the end of the statement.

Funding

This work was supported by the R + D project NVAL24/10, funded by Instituto de Investigación Valdecilla (IDIVAL).

Data availability

We deposited the genome sequences of our strains in the NCBI GenBank under Bioproject PRJNA1230562.

Declarations

Ethics approval and consent for publication

This manuscript follows the ethical principles included in the Declaration of Helsinki. The publication of clinical data and the procurement of clinical isolates used in the study was approved by the Cantabria’s Committee on Ethics in Health Medicine Research with code: 2024.280. Cantabria’s Committee on Ethics in Health Medicine Research waived from the requirement for written informed consent as this was a retrospective, observational study, where patients were no longer available at the time of the study and were difficult to locate.

Competing interests

The authors declare no competing interests.