Molecular Epidemiology of Colonizing and Infecting Isolates of Klebsiella pneumoniae

Abstract

Klebsiella pneumoniae is among the most common causes of hospital-acquired infections and has emerged as an urgent threat to public health due to carbapenem antimicrobial resistance. K. pneumoniae commonly colonizes hospitalized patients and causes extraintestinal infections such as urinary tract infection, bloodstream infection (septicemia), and pneumonia. If colonization is an intermediate step before infection, then detection and characterization of colonizing isolates could enable strategies to prevent or empirically treat K. pneumoniae infections in hospitalized patients. However, the strength of the association between colonization and infection is unclear. To test the hypothesis that hospitalized patients become infected with their colonizing strain, 1,765 patients were screened for rectal colonization with K. pneumoniae, and extraintestinal isolates from these same patients were collected over a 3-month period in a cohort study design. The overall colonization prevalence was 23.0%. After adjustment for other patient factors, colonization was significantly associated with subsequent infection: 21 of 406 (5.2%) colonized patients later had extraintestinal infection, compared to 18 of 1,359 (1.3%) noncolonized patients (adjusted odds ratio [OR], 4.01; 95% confidence interval, 2.08 to 7.73; P < 0.001). Despite a high diversity of colonizing isolates, 7/7 respiratory, 4/4 urinary, and 2/5 bloodstream isolates from colonized patients matched the patient corresponding rectal swab isolates, based on wzi capsular typing, multilocus sequence typing (MLST), and whole-genome sequence analysis. These results suggest that K. pneumoniae colonization is directly associated with progression to extraintestinal infection. IMPORTANCE K. pneumoniae commonly infects hospitalized patients, and these infections are increasingly resistant to carbapenems, the antibiotics of last resort for life-threatening bacterial infections. To prevent and treat these infections, we must better understand how K. pneumoniae causes disease and discover new ways to predict and detect infections. This study demonstrates that colonization with K. pneumoniae in the intestinal tract is strongly linked to subsequent infection. This finding helps to identify a potential time frame and possible approach for intervention: the colonizing strain from a patient could be isolated as part of a risk assessment, and antibiotic susceptibility testing could guide empirical therapy if the patient becomes acutely ill.

Keywords: Klebsiella; MLST; cgMLST; colonization; infection; pneumonia; whole-genome sequencing; wzi.

FIG 1

Study population. Adult patients in the University of Michigan Health System’s ICU and adult hematology/oncology patients were screened for colonization and extraintestinal infection with K. pneumoniae between July and October 2014 (n = 1,765), divided into “infected” and “not infected” groups, and further divided into “colonized” and “not colonized.” The number of infections by body site are shown in boxes; one colonized patient met case definitions for both pneumonia and UTI.

FIG 2

Receiver operator characteristic curve for a multivariable model of risk factors for clinical infection. Multiple logistic regression of K. pneumoniae infection was used to generate a predictive model using five patient variables (Table 4). Bars and shaded areas of ROC curves represent bootstrapped 95% confidence intervals (10,000 replicates) for specificity at each level of sensitivity (AUC, 0.78; 95% CI, 0.72 to 0.84).

FIG 3

Phylogenetic tree for wzi sequence of patient rectal swab isolates. Unique patients (P) are numbered (P1 to P40). A rectal swab (S) isolate is indicated after the patient number and immediately before the isolate number (e.g., S463 is stool isolate number 463). The isolate wzi type is indicated, and novel alleles are designated unknown by UK. A total of 110 rectal swab isolates from 40 unique patients were tested for strain type using wzi gene sequencing. A total of 43 different wzi types of strains were identified. Rectal swab isolates for patients with K. pneumoniae colonization prior to infection were all included in the analysis (P1 to P16; colored font). The scale bar represents the amount of genetic change; 0.01 equals 1 change per 100 nucleotide sites. The numbers next to each node are the percentage of iterations that recovered the same node.

FIG 4

Core genome similarity between infecting and colonizing strains within patients. Patients who had concordant colonizing and infecting isolates based on wzi sequencing were further analyzed by WGS and core genome MLST and are represented by a unweighted pair group method using average linkages dendrogram along with isolate number, MLST type (ST), wzi type, and body site of culture (source). Each color represents an individual patient. Isolates 463 and 1946 (patient 1) were discordant by wzi and were included as a control.