The retrospective study was conducted in a teaching hospital of Tongji University (Shanghai, China). From January 2012 to November 2014. All CRKP isolates were collected from the database of the Microbiology Service of the Shanghai Tenth People's Hospital of Tongji University. The clinical records of patients with CRKP isolates were reviewed. The enrollment criteria were as follows: i) Clinical evidence of sepsis, severe sepsis or septic shock due to CRKP; ii) meropenem or imipenem were used within 72 h after CRKP cultured from urine, blood, phlegm or drainage liquid; iii) the administration of meropenem therapy lasted for at least 72 h; and iv) >18 years old.

Admission to intensive care unit (ICU) was defined as a stay of 24 h before or after diagnosis of the severe sepsis. Predictors of the primary outcomes included age, sex, underlying diseases [diabetes mellitus, history of chronic obstructive pulmonary disease (COPD), heart failure, hepatic failure, renal failure, gastrointestinal surgery, solid malignancies, hematologic malignancies], special treatments (central venous catheter, peripheral catheter, foley catheter, Nasogastric tube, mechanical ventilation ≥3 days, gastric acidity-lowering agents, immunosuppressive therapy, total parenteral nutrition), prior use of antibiotics and bacteremia. Clinical outcome contained length of ICU stay, mortality or cure, and duration of mechanical ventilation. The patients with recurrent CRKP isolated from the blood or urine were defined as microbiological failure at least 7 days after their index culture. Patients who had a negative culture were defined as treatment success.

Quantitative variables are expressed as the mean ± standard deviation and qualitative variables are depicted as percentages of the group to which they belonged. Differences between patients treated with fosfomycin and meropenem and other antibiotics were analyzed by univariate logistic regression. Risk factors correlated with clinical outcome and microbiological failure was analyzed using stepwise multiple logistic regression analyses. The odds ratio (OR) was calculated with 95% confidence interval (CI). All statistical tests were two-tailed and utilized a 0.05 significance level. Analyses were performed using SPSS software, version 20.0 (IBM SPSS, Armonk, NY, USA).

A total of 104 unique cases of bacteremia with CRKP were identified during the experimental period. The demographics and clinical characteristics of the cases are shown in Table I. The ages of the 104 patients, 79 male and 25 female, ranged from 28 to 95 years, with a median of 67.2 years. The majority of the infections (84; 80.8%) were hospital acquired. All but one patient had been admitted to the hospital within a year prior to the episode of bacteremia, with a majority of them (87; 83.7%) having been admitted to an ICU. Patients that received ineffective empirical antimicrobial therapy before the susceptibility results became available are in the majority and 80 patients (76.9%) received mechanical ventilation. The majority of patients had underlying diseases, had undergone invasive procedures, and had been admitted to ICUs or received medical services.

Colistin exerted the highest susceptibility rate with 93.3% (Table II). Tigecycline and minocyline were active against 68 (65.4%) and 79 (76.5%) isolates, respectively. With regard to fosfomycin, 40 isolates (38.5%) were susceptible, 54 (51.9%) showed intermediate susceptibility and 10 (9.6%) were resistant. A less marked susceptibility to amikacin (28, 26.9%) and gentamicin (14, 13.5%) was observed.

All patients received targeted antibiotic treatment. Targeted treatment was optimal in 78 patients (75.0%). The targeted antibiotics used were shown in Table III. A total of 10 patients were infected with a fosfomycin-resistant CRKP strain. This combination treatment group consisted of 24 patients (23.1%) that were administered a dose of 12 mg/24 h and were not administered fosfomycin as monotherapy. In the fosfomycin combinational therapy group, 16 patients were administered meropenem (1 g every 8 h) following with fosfomycin with the other 8 patients received fosfomycin, meropenem, tigecycline or (and) minocyline. Those patients received fosfomycin combination were less likely to fail therapy (OR: 4.71, 95% CI: 1.03–21.65, P=0.030). However, no difference between length of ICU stays and duration of mechanical ventilation was observed in these two groups (Table I). The primary characteristics of the patients who received fosfomycin combinational therapy are shown in Table IV. According to the result, the number of patients that underwent prior use of minoglycosides and carbapenems differed between the two groups. Compared with other regimens (10.2±11.0 vs. 13.0±11.2), fosfomycin combination group tends to have shorter duration of mechanical ventilation, however the difference did not display statistical significance. An interesting thing is that the age of fosfomycin combinational therapy group is older than other treatments group (68.7±14.6 vs. 62.0±18.3, P=0.065), which might give us a hint that fosfomycin combinational therapy is safer and more suitable to be applicable to aged people. The fosfomycin combinational therapy group has an overall mortality of 8.3% with one case died of pneumonia and another died of the infection of the central nervous system (Table V).

Taking the difference in prior use of aminoglycosides into consideration, the effect of fosfomycin in combination with meropenem and with other regimens treatment was compared (Table VI). This combinational therapy group has 16 patients. Compared with other therapy, a significant difference was observed (Table VI, P=0.010) when compared with treatment by using 3rd/4th generation (monotherapy or combinational therapy).

Differences between cases of mortality and survival was observed using univariate analysis (Table I). Peripheral catheter and history of COPD have a negative influence on survival rates (P=0.01 and P=0.016). Meanwhile, compared with female patients, male patients have a higher mortality. Fosfomycin combinational therapy (7.7 vs. 24.6%) was associated with lower mortality (Table I). Compared with other targeted therapy, target treatment with fosfomycin exerted preponderance (P=0.034, OR: 4.71 [1.03–21.65]). The survival analysis performed using Kaplan-Meier curves showed that patients treated with fosfomycin combinational therapy had higher survival rates at 45 days after diagnosis (Fig. 1, log-rank test t=3.96, P=0.047). However, in a multivariate analysis, no correlation between fosfomycin combination and outcome was observed (Table VII, P=0.141).