The basic characteristics of KP-infection patients
A total of 425 pediatric patients met the diagnosis criteria of invasive KP infections based on the hospital microbiology laboratory records, of whom, 248 (58.4%) neonates and 177 (41.6%) were non-neonates on admission. Finally, 324 (76.2%) patients were included in this study and 101 (23.8%) were excluded due to unavailability of medical records (99 patients) and incomplete clinical data collection (2 patients) (Fig. 1). The patients were categorized into three groups for further analysis (Fig. 1).
Of the 324 invasive KP-infected inpatients, 299 (92.3%) had KP isolated from blood, 11 (3.4%) had KP isolated from CSF and 14 (4.3%) had KP isolated from both blood and CSF; 178 (54.9%) were from rural area; 187 (57.7%) were male; the median age was 0.8 months (IQR 0.2-6.375, range: 1 day-17 years) and the age distribution was as followings: 188 (58%) neonates, 34 (10.5%) aged 29 days-3 months, 32 (9.9%) aged 3–11 months, 40 (12.3%) aged 1–5 years, 30 (9.3%) aged > 5 years. Of 188 neonates, 131 (69.7%) were preterm including 13 (9.9%) at < 28 weeks, 75 (57.3%) at 28–32 weeks and 43 (32.8%) at 32-37weeks, according to gestational age.
On the clinical grounds and case definition (Table Ⅰ), HAI and CAI were considered in 275 (84.9%) and 49 (15.1%) patients, respectively. CRKP infections occurred in 116 (35.8%) patients. Two hundred and fifty-nine (79.9%) patients had the accompanying organ infections, including pneumonia (139, 42.9%), meningitis (67, 18.5%), intra-abdominal infections (46, 14.2%), upper urinary tract infections (18, 5.6%) and osteomyelitis (5, 1.5%); 16 patients had 2 sites of infection. Nine (2.8%) patients died in hospital and 34 (10.5%) patients were discharged hopelessly at the request of parents.
As shown in Table Ⅱ, univariate analysis showed that invasive mechanical ventilation, any organ dysfunction, septic shock and accompanying pneumonia were the significant risk factors for poor outcomes while patient receiving ≥ 1 active antibiotic therapy had lower risk of poor outcomes. The poor outcomes were similar between CRKP and CSKP infections (13.8% versus 13.0%, P > 0.05). Multiple regression analysis showed that organ dysfunction was an independent risk factor for poor outcomes (OR:2.88, [95% CI: 1.25–6.60], P < 0.05).
Comparison of clinical features in patients with CAI and HAI
Compared to CAI patients (shown in Table Ⅰ), the percentage of patients with underlying medical conditions was significantly greater in HAI patients, including hematologic malignancies, necrotizing enterocolitis (NEC) and prematurity (16.4% versus 2%, 9.8% versus 0%, 46.9% versus 4.1%, respectively; P < 0.05); the percentage of patients using peripherally inserted central catheter (PICC), invasive mechanical ventilation and blood products was significantly greater in HAI patients (29.8% versus 2%, 35.6% versus 2%, 85.5% versus 36.7%, respectively P < 0.01). HAI patients were more frequently accompanied by meningitis than CAI patients (23.3% versus 6.1%, P < 0.01) while CAI patients were more frequently accompanied by upper urinary tract infections (1.1% versus 30.6%, P < 0.01). Of note, the percentage of CRKP strains was much higher in HAI patients than in CAI patients (40.4% versus 10.2%. P < 0.01). The proportion of poor outcomes were similar in HAI and CAI patients (14.2% versus 8.2%, P > 0.05).
Multiple regression analysis showed that hematologic malignancies (odds ratio (OR:15.52, 95% CI:1.89-127.14), prematurity (OR:37.07, 95% CI:8.29-165.84), invasive mechanical ventilation (OR:13.09, 95% CI: 1.66-103.56) were independent risk factors for HAI (Table Ⅲ).
Comparison of clinical features in patients with CRKP and CSKP infections
Compared to CSKP-infected patients (shown in Table Ⅰ), more CRKP-infected patients were premature and from the rural area (49.1% versus 35.6%, 66.4% versus 48.6%, respectively; P < 0.05,), spent longer time and higher medical expenses on hospitalization (37.2 versus 44.4 days, 78508.8 versus 57919 RMB Yuan, respectively; P < 0.01), received therapies of mechanical ventilation and blood products (58.6% versus 37.5%, 85.3% versus 74%, respectively P < 0.05), and received more number and longer duration of antibiotics (3.7 versus 4.8, 39.8 versus 27 days, respectively; P < 0.01). Less CRKP-infected patients received empirical active antibiotics and definitive target treatment than CSKP-infected patients (19.8% versus 77.4%, 37.1% versus 76.9%, respectively; P < 0.01). The poor outcomes were similar between CSKP-infected and CRKP-infected patients (13.8% versus 13%, P > 0.05).
Multiple regression analysis showed that residing in rural area (OR: 2.05, 95% CI:1.22–3.43, P < 0.05), invasive mechanical ventilation (OR:2.06, 95% CI: 1.13–3.76, P < 0.05), previous antibiotic therapy (OR: 2.07, 95% CI:1.17–3.68, P < 0.05) and previous hospital stay (OR: 2.18, 95% CI:1.18–4.03, P < 0.05) were independent risk factors for CRKP infections (Table Ⅳ). For neonates, invasive mechanical ventilation (OR:2.19, 95% CI: 1.06–4.53, P < 0.05) and previous hospital stays (OR:2.26, [95% CI: 1.02–4.99], P < 0.05) were significant risk factors for CRKP infections. (Table Ⅴ)
Comparison between neonates and non-neonates (children > 28 days of age) with KP infections
Compared to children > 28 days of age (Table Ⅵ), the prevalence of CRKP was significantly higher in neonatal patients (39.9% versus 30.2%,P < 0.05), and the occurrence of HAI was significantly higher in neonates (91.5% versus 55.9%, P < 0.05); neonates are more likely to develop severe KP-associated meningitis and organ dysfunction (22.7% versus 11%, 28.7% versus 14%; P < 0.01), and receive longer course of antibiotics and invasive mechanical ventilation (34 days versus 28.3 days, 39.4% versus 19.6%, respectively; P < 0.01). The overall length of hospital stay and medical expenses of hospitalization were significantly higher in neonates than non-neonates (38.5 versus 35.3 days, 72426.5 versus 56222.9 RMB, respectively; P < 0.01). The poor outcomes were similar between neonates and non-neonates (13.3% versus 13.2%, P > 0.05).
Antimicrobial susceptibility patterns
As shown in Table Ⅶ, KP showed high percentage of resistance to clinical important antibiotics usually recommended for treatment of Enterobacteriaceae infection, such as third-generation cephalosporins, cefepime, carbapenems. KP almost remained sensitive to tigecycline and polymyxin B, and also showed high frequency of sensitivity to amikacin, amikacin, ciprofloxacin, and levofloxacin.
Of note, the healthcare-associated KP (HA-KP) strains showed higher frequency of resistance to all tested antibiotics than the community-acquired KP (CA-KP) stains. The CA-KP stains displayed low frequency of resistance to gentamicin, amikacin, ciprofloxacin, and levofloxacin (2.2%-8.7%). In terms of age groups, neonates displayed higher frequency of resistance to all tested antibiotics, especially to fourth- and third-generation cephalosporins and amikacin. CRKP also displayed high frequency of resistance to cephalosporins and piperacillin/tazobactam (90.4%-100%) than CSKP. CSKP were all sensitive to amikacin and showed low frequency of resistance to ciprofloxacin and levofloxacin (5.9% and 2.1%).