Risk factors, resistance profiles, and clinical outcomes of Klebsiella pneumoniae infection in cancer patients: a retrospective study.

BACKGROUND: Cancer patients are highly susceptible to Klebsiella pneumoniae (KP) infection, yet integrated analyses considering infection-site resistance patterns, cancer-type heterogeneity, and host risk factors remain limited. METHODS: We retrospectively analyzed 300 adult cancer patients treated between October 2023 and November 2024, including 153 with KP infection and 147 uninfected controls. Data on demographics, comorbidities, invasive procedures, laboratory findings, antimicrobial resistance, treatment patterns, and outcomes were collected. Independent risk factors were identified using multivariable logistic regression, and a predictive nomogram was developed and validated. RESULTS: The respiratory tract was the most common infection site (60.8%), followed by urinary tract (14.4%) and bloodstream (10.4%). Urinary isolates exhibited the highest rates of multidrug-resistant KP (MDR-KP, 69.6%) and extended-spectrum β-lactamase-producing KP (ESBL-KP, 60.9%). Multivariate analysis identified hypoproteinemia (odds ratio [OR] = 2.233, 95% confidence interval [CI] 1.222-4.077, P = 0.009), percutaneous catheter drainage (OR = 1.995, 95% CI 1.050-3.790, P = 0.035), and radiotherapy (OR = 3.556, 95% CI 1.507-8.392, P = 0.004) as independent factors associated with KP infection. Recent hospitalization was independently associated with both MDR-KP (OR = 3.257, 95% CI 1.227-8.649, P = 0.018) and ESBL-KP (OR = 3.101, 95% CI 1.151-8.355, P = 0.025). A nomogram incorporating these predictors demonstrated strong discriminative ability (area under the curve = 0.702) and clinical utility. CONCLUSION: Statistical models indicate that KP infection in cancer patients is likely linked to host immunity, invasive interventions, and treatment exposures; the nomogram can assist in early risk stratification and facilitate tailored stewardship.