The study population included 189 patients greater than 14 years of age with a positive CRE culture and a clinically established infection. The median patient age was 62.8 years and 54.0% were male (Table 1). Just over one-fifth (20.1%) of the participants had heart disease and 42.3% had diabetes mellitus. Almost one-third (32.8%) of the patients had a non-hematological malignancy and 19.6% had a cerebral vascular accident comorbidity. The median Charlson comorbidity index was six at baseline. The APACHE II median score for patients who required ICU admission was 19 (n = 112) while the Pitt bacteremia median score for patients with bacteremia was two (n = 110) (Table 1).
Table 1 Demographic and baseline clinical characteristics of patients with CRE infection (n = 189) Almost half of the patients had either hospital-acquired pneumonia (HAP) or ventilator-associated pneumonia (VAP) (23.8%) or a complicated urinary tract infection (UTI) (23.8%) caused by CRE (Table 1). The majority of organisms isolated were Klebsiella pneumoniae (87.3%). Seventy-seven patients (40.7%) had CRE bacteremia and approximately one-third (30.2%) of patients were in the ICU at the time they received a positive CRE culture (Table 1).
For the isolates tested, the median MIC was greatest for meropenem with a value of 16 followed by imipenem with a value of 8. Of the 143 isolates tested for CAZ-AVI, 103 (72.0%) were susceptible, 35 (24.5%) were resistant, and 5 (3.5%) were intermediate at baseline. For the isolates resistant to CAZ-AVI, the molecular typing was KPC (n = 1), NDM (n = 24), and OXA-48 (n = 15) with some overlap between KPC and NDM being found in 1 patient and NDM and OXA-48 being found in 8 patients. The median MIC was smallest for CAZ-AVI with a value of 2 for the isolates tested against this antibiotic. Molecular typing revealed over two-thirds (69.3%) of patients were infected with OXA-48 CRE (Table 1).
Overall, the study included 114 patients from western region hospitals, 41 patients from the central region of the Kingdom, and 34 patients from the Eastern province (Fig. 1). While 78.9% of patients in the western region had OXA_48 molecular typing identified in their isolates, 43.9% of patients in the central region and 67.7% of patients in the eastern region had OXA_48 identified in their isolates. A similar percentage of the patient isolates were identified as NDM in the western (14.9%) and central (14.6%) regions compared to 26.5% of patient isolates being identified as NDM in the eastern region. Almost one-third of isolates (31.7%) from patients in the central region had no molecular typing detected (Fig. 1).
While 69.3% of patients took a carbapenem as part of their treatment regimen, 47.1% of patients received CAZ-AVI (Table 2). Approximately one-fifth (20.1%) of patients received an aminoglycoside at some point during their treatment regimen (Table 2). There were 100 patients (52.9%) with a clinical cure and 57 patients (30.2%) had died within 30 days (Table 3). Over one-third (36.5%) of patients had acute kidney injury and 23 patients (12.2%) relapsed within 30 days (Table 3).
Table 2 Treatment regimen for patients with CRE infection Table 3 Outcomes of patients with CRE infections (n = 189) The median age of patients who died within 30 days at 64.7 years was significantly different than the median age, 60.3 years, in the group of patients who did not die within 30 days (Kruskal–Wallis Test = 4.8, df = 1, p = 0.03). Similarly, the association of Charlson comorbidity index with 30 day mortality was significant with a median of 6.0 in the group of patients who died within 30 days compared to a median of 5.5 in the group of patients who did not die within 30 days (Kruskal–Wallis Test = 9.1, df = 1, p = 0.003). The median values of Pitt bacteremia were also significantly different between the two groups with a value of 6.0 in the group of patients who died compared to a median value of 1.0 in the group of patients who did not die within 30 days (Kruskal–Wallis Test = 15.3, df = 1, p < 0.0001).
There was a statistically significant difference between the patients who had a HAP/VAP CRE infection with regards to 30 day all-cause mortality (Chi-Square = 5.7, df = 1, p = 0.02). While 35.1% of the patients who died within 30 days had a HAP/VAP infection, 18.9% of patients who did not die within 30 days had a HAP/VAP infection. The association between having CRE bacteremia and 30 day all-cause mortality was also statistically significant with 57.9% of patients who died also having CRE bacteremia compared to 33.3% of patients who did not die having CRE bacteremia (Chi-Square = 9.9, df = 1, p = 0.002). Presenting with the comorbid condition of Diabetes Mellitus was also significantly associated with 30 day all-cause mortality with 54.4% of patients who died having this condition compared to 37.1% of patients who did not die within 30 days having this condition (Chi-Square = 4.9, df = 1, p = 0.03).
There was no significant association identified between CRE treatment with CAZ-AVI and 30 day all-cause mortality (Chi-Square = 0.8, df = 1, p = 0.37). Twenty four (27.0%) of the patients who took CAZ-AVI ended up dying within 30 days of all-cause mortality. Approximately one third (n = 33, 33.0%) of patients who did not take CAZ-AVI (n = 97) ended up dying within 30 days of all-cause mortality.
In further looking at the association of CAZ-AVI treatment with 30 day all-cause mortality, among the patients with isolates that tested sensitive to CAZ-AVI and who took CAZ-AVI as part of their treatment regimen compared to the rest of the patients, there was no significant relationship. The association between those who tested sensitive compared to those with resistant or intermediate sensitivity results to CAZ-AVI for the patients who took CAZ-AVI was also not significant with regards to 30 day all-cause mortality. In addition, taking CAZ-AVI treatment did not significantly predict whether or not patients had a clinical cure [OR and 95% CI = 1.52 (0.86–2.71), p = 0.15]. The outcome of acute kidney injury was not significantly associated with CAZ-AVI treatment [OR and 95% CI 1.15 (0.63–2.08), p = 0.65].
Patients with CRE bacteremia (n = 77) had a median age of 63.6 years, a median Charlson comorbidity index of 6.0, and a median Pitt bacteremia score of 3.0. Among the bacteremic patients, the association between having CAZ-AVI as part of the treatment regimen and 30 day all-cause mortality was significant [odds ratio (OR) and 95% CI = 0.32 (0.12–0.81), p = 0.02] showing a protective effect of the treatment in univariate analysis. After controlling for age, Charlson comorbidity index, and Pitt bacteremia score the association between treatment with CAZ-AVI and 30 day all-cause mortality was no longer significant [AOR and 95% CI = 0.33 (0.10–1.06), p = 0.06]. After adjusting for these predictors, a one-unit increase in Pitt bacteremia score was associated with an increase of 29.0% in the odds of 30 day all-cause mortality [AOR and 95% CI = 1.29 (1.09–1.52), p = 0.004].
Univariate logistic models were constructed to predict the outcome of 30 day all-cause mortality based on demographic and clinical patient characteristics (Table 4). Univariate analysis revealed that a one-unit increase in the Charlson comorbidity index was associated with a 20.0% increase in the odds of dying within 30 days from all causes [OR and 95% CI = 1.20 (1.07–1.33), p = 0.001] and patients with a HAP/VAP infection had an increased odds by a factor of 2.3 of dying within 30 days from all causes [OR and 95%CI = 2.31 (1.15–4.64), p = 0.02]. For patients with a diagnosis of diabetes mellitus at baseline, the odds of dying within 30 days from all-causes increased by a factor of 2.2 times compared to patients without this diagnosis [OR and 95% CI = 2.02 (1.08–3.79), p = 0.03]. All predictor variables with a p < 0.1 were included in the multivariable logistic model looking at the association of clinical characteristics and 30 day all-cause mortality (Tables 4, 5). After adjusting for all predictors with a p < 0.1 from univariate analyses, patients with CRE bacteremia had an increased odds by 281% of dying within 30 days from all causes [AOR and 95% CI = 2.81 (1.26–6.24), p = 0.01]. (Table 5).
Table 4 Univariate analysis of patient characteristics associated with 30 day all-cause mortality Table 5 Multivariable association of characteristics with 30 day all-cause mortality (n = 189)