Carbapenem-Resistant Klebsiella pneumoniae Urinary Tract Infection following Solid Organ Transplantation

DISCUSSION

To our knowledge, the present study is the largest detailed clinical characterization and comparison of OTRs with UTIs due to CRKP, ESBL-producing K. pneumoniae, or susceptible K. pneumoniae. We observed that UTIs caused by CRKP were associated with greater odds of microbiological failure. We documented that UTIs due to CRKP or ESBL-producing K. pneumoniae occurred significantly earlier after transplantation than those due to susceptible K. pneumoniae and were associated with a urinary catheter at diagnosis, a ≥24-hour stay in the ICU either before or after development of the UTI, and longer LOS prior to the UTI. These data support the idea that, during the study period, CRKP was largely a nosocomial pathogen. We further observed a trend toward an increased likelihood of patients with CRKP experiencing relapse within 30 days and continued fever or urgency, frequency, dysuria, suprapubic tenderness, or costovertebral angle pain or tenderness 72 h after receiving adequate antimicrobial therapy. The mean time to initiation of that therapy was significantly longer for CRKP.

UTI occurs more frequently in OTRs than in the nontransplant population. For example, UTI, estimated to be associated with approximately 44 to 47% of all infectious complications (5), is the most common infection following kidney transplantation, but published studies of UTI due to CRKP in OTRs are lacking. Simkins and colleagues compared 9 kidney transplant recipients with UTIs due to CRKP and 34 with carbapenem-susceptible K. pneumoniae (6). They observed a trend toward more bacteremia among cases of CRKP, as well as a significantly higher mortality rate at 6.5 months. Unfortunately, other investigations into CRKP among OTRs excluded UTIs (7–10). Still others included both OTRs and nontransplant patients but focused on a particular therapy (11), bacteremia (12, 13), or factors associated with the development of UTIs among patients with asymptomatic bacteriuria (14). Extrapolating from data for nontransplant patients is problematic. It has been shown that, among mostly (81%) nontransplant patients with UTIs due to CRKP, microbiological failure occurred in only 5 cases (24%) and bacteremia in 3 (14%) (15). In contrast, we observed microbiological failure for 45% of OTRs and bacteremic UTI due to CRKP for 32%. In addition, it was shown that aminoglycosides were associated with clearance of CRKP bacteriuria among nontransplant patients (16), but our data demonstrated that 3/6 aminoglycoside-exposed OTRs (50%) experienced microbiological failure. We hypothesize that immunosuppressive medication enhances the importance of early active antimicrobial therapy and accounts in part for the increased rates of microbiological failure and bacteremia.

Our results are generally in agreement with the other very limited published studies. In Brazil, Argentina, and Israel, 12 OTRs with UTIs due to CRKP have been described, 6 (50%) of whom were bacteremic (17–19). Mortality rates ranged from 0 to 33%. Finally, in a recent study of patients from the United States, there were 11 OTRs with UTIs due to CRKP (14). While detailed clinical characterizations were not provided, multivariate logistic regression analyses identified solid organ transplantation as independently associated with UTI among all patients with CRKP bacteriuria (OR, 4.50 [95% CI, 1.39 to 14.52]).

In the present study, we report an additional 22 OTRs with UTIs due to CRKP, with ESBL-producing K. pneumoniae and susceptible K. pneumoniae for comparison. As shown, many features of UTIs due to ESBL-producing K. pneumoniae were similar to those for CRKP; however, important differences emerged, particularly in the increased odds of microbiological failure associated with CRKP. Failure to include ESBL-producing K. pneumoniae would have placed the findings in an improper context, and the results would have been much less informative. Although our study provides some advantages, the observational design is limited by unmeasured confounders, such as the presence of ureteral stents and confounding by indication. Therefore, some of our observations should be interpreted cautiously. These data reflect the biases of the clinicians at our institution. For example, providers more often repeated urine cultures for OTRs with UTIs due to CRKP or ESBL-producing K. pneumoniae, which could bias the results with regard to microbiological failure, in comparison to susceptible K. pneumoniae. Finally, ESBL-producing K. pneumoniae and carbapenem resistance were defined phenotypically. During the study period, testing for K. pneumoniae carbapenemase (KPC) production through detection of bla_KPC was not performed; however, subsequent PCR testing of the isolates with positive results from the modified Hodge test revealed the overwhelming majority to be bla_KPC positive. While the phenotypic definitions may be somewhat limited, results reflect the information available to providers caring for these patients in a real-world setting. Therefore, our data incorporating these definitions remain largely informative and relevant.

In conclusion, CRKP is associated with microbiological failure among OTRs with UTIs and may be an important predictor of clinical outcomes. There are few therapeutic options for CRKP, and use of the available agents is limited by the parenteral route of administration and toxicity, particularly among patients with recent or ongoing organ failure syndromes, which are frequently encountered among OTRs. This underscores the importance of infection prevention. Additional studies with OTRs, with greater patient numbers and an emphasis on determining modifiable risk factors for the development of infection, are needed. These will permit discovery of significant differences in clinically meaningful endpoints associated with carbapenem resistance, which would be invaluable to clinicians caring for OTRs.