Secondary outcomes:
Out of the 205 patients with isolated Klebsiella organisms in the ICU during the study duration, 122 (59.5 %) were carbapenem-resistant (CRK). Those CRK were resistant to imipenem in 97 patients, to meropenem in 62 patients, or resistant to both in 51 patients.
The pre-intervention CRK isolates was reported in 52 isolates (85.2 %) compared to 70 isolates (48.6 %) in the post-intervention period. These differences were statistically significant with OR (95% CI) of 0.164 (0.075-0.357), P < 0.001.
Among the 61 patients in the pre-intervention period, no patients were in grade 1 (0 %), while 9 (14.8 %) and 52 patients (85.2 %) had grade 2 and 3 resistance respectively compared to 13 (9 %), 61 (42.4 %) and 70 (48.6 %) of the 144 of the post-intervention periods had resistance of grade 1, 2, and 3 respectively (P < 0.001). This pattern showed a decreased prevalence of CRK (grade 3) at the expense of increased grades 1 and 2 Klebsiella species.
One-way ANOVA (Levene’s Test of equality of error variances p-value > 0.05) showed that the difference in carbapenem consumption reported with the different Klebsiella grades was significant (F=4.8, P=0.009). It was 31.9 ± 3 DDD / 1000 patient-days with grade 1 compared to 31.7 ± 10.1 and 36.5 ± 11.8 DDD / 1000 patient-days with grades 2 and 3 respectively. Moreover, post-hoc analysis using Turkey’s test for pairwise comparison found a significant difference between grade 2 and 3 (P=0.01) (Figure 2).
Figure 2: The relationship between Klebsiella resistance grades and carbapenems consumption.
Two ordinal logistic regression models were used. The first model with a fit parallel line test (χ2 = 1.97, P = 0.16) showed that Klebsiella species were more likely to be in a lower category of resistance significantly after implementation of the ASP with OR (95 % CI) of 6.3 (2.88 – 13.73). The second model with a fit parallel line test (χ2 = 0.275, P = 0.6) showed that that Klebsiella species were more likely be in a lower category of resistance with a decrease in carbapenems use with OR (95 % CI) of 1.04 (1.012 – 1.07).
4. DISCUSSIONS:
Multidrug-resistant bacteria are shown to be caused by antibiotics abuse in the healthcare facilities (Cantón, Horcajada, Oliver, Garbajosa, & Vila, 2013) especially in ICUs (Prabaker & Weinstein, 2011). We hypothesized that the application of ASP within the ICU including the preauthorization of restricted antibiotics list, the timeout for de-escalation, and the timeout for early discontinuation with other different elements can decrease the consumption of different antibiotics especially carbapenems, and consequently decrease the prevalence of the CRK species. We conducted an observational study to measure the carbapenem consumption within the ICU of a tertiary hospital in an Egyptian rural area after the implementation of ASP elements and compared this consumption with that prior to the program implementation. We also evaluated the impact of the carbapenem consumption on theKlebsiella species resistance pattern.
The implementation of the ASP in our study significantly decreased the carbapenem utilization, where it caused a decrease in the carbapenem use by 31.6 % compared to 22% in another study (Elligsen et al., 2012). A previously reported stewardship program decreased carbapenem utilization from 300 to 20 DDD / month (Simmons, Sherman, & Crosmun, 2016), which is not like our results that could be because of the different units used for carbapenem consumption. Another study found that the ASP decreased the consumption of meropenem from 90.5 to 24.96 and imipenem from 6.5 to 2.3 in terms of DDD / 1000 patients-days (Jaggi, Sissodia, & Sharma, 2012).
Other impacts of ASP were studied, and it was concluded that the application of the program caused a significant decrease in the inappropriate antibiotic use from 18.5 % to 11.4 % (Maeda et al., 2016) and the implementation of antibiotic restriction policy decreased the total broad-spectrum antibiotic days. (Rimawi, Mazer, Siraj, Gooch, & Cook, 2013). Moreover, the prospective audit and feedback elements of the ASP were seen to be of utmost importance in applying of the program in many studies (Lesprit, Landelle, & Brun-Buisson, 2013; D. H. Solomon et al., 2001). One study applied an easy-to-use electronic preauthorization system within ASP that was seen to significantly decrease the cephalosporins use by 39 % (Buising et al., 2008). Other earlier programs that restricted the use of cephalosporins caused an 18 % decrease in hospital prescription of cephalosporins (Hanberger, Skoog, Ternhag, & Giske, 2014).
Appropriate antibiotics selection and time to de-escalation from carbapenems or antipseudomonal antibiotics was studied after ASP implementation and it was found that ASP significantly improved time to de-escalation and appropriateness of antibiotics (Bookstaver et al., 2017). We didn’t however analyze the impact of the different program elements separately due to the small sample size. The decrease in carbapenem consumption in our study might indicate compliance to restricted antibiotics policy, shorter time to antibiotics de-escalation, appropriate duration, and optimum dosing.
We reported a 59.5 % prevalence of CRK in our study. Other studies in Egypt reported similar prevalence ranging from 44 % up to more than 63 % (Amer et al., 2016; El Kholy & El Manakhly, 2018; Metwally et al., 2013). Additionally, out of 80 Klebsiella pneumoniae  isolates obtained from hospitalized patients at Menoufia University Hospitals, the resistance to imipenem and meropenem was observed in 62.5 and 56.2 respectively(Melake et al., 2016).
We reported a significant decrease in the prevalence of the CRK after the implementation of the ASP interventions. Viñau Lopez et al. found a decrease in resistance of Klebsiella pneumoniae after ASP implementation in a tertiary hospital from 46 % to 38 % (Jaggi et al., 2012). Other studies reported that the applying of ASP with a shorter antibiotic course was shown to decrease antibiotic-resistant superinfection from 38 % to 14 % (Srinivasan, 2017). Others showed that interventions that decreased excessive antibiotic use, decreased the antibiotic resistance pattern, and even improve clinical outcomes (Davey et al., 2013). Despite that Elligsen et al showed that the ASP interventions decreased meropenem resistance, they didn’t show a significant difference in the clinical outcome in terms of mortality or length of stay (Elligsen et al., 2012).
We also found that higherklebsiella resistance grades were associated with higher carbapenem consumption. Horikoshi et al showed a positive correlation between days on carbapenems and carbapenems resistance in the pediatric population (Horikoshi et al., 2017). Another case-control study involving Klebsiella pneumoniae bloodstream infection found that carbapenem use caused 9.98 fold increase in the odds of acquiring carbapenem resistance (Yuan et al., 2020). Carbapenem exposure within one month before the onset of bloodstream infection was found by other authors to be the only risk factor for developing CRK infection (Liu et al., 2019).
Despite that the reduced antibiotic consumption could be the direct outcome of applying ASP, its long-term impact might include several clinical and economic impacts. Mortality reduction and decrease hospital length of stay as outcomes to ASP had been also studied. The ASP implementation caused up to a 54 % reduction in mortality with an OR of 2.17 (Conway et al., 2017) and decreased length of hospitalization (Horikoshi et al., 2017). These results were, however, not reproduced in a meta-analysis including 11 studies (Lindsay et al., 2019).
Our study was limited by the small sample size. We did not study the impact of the program on clinical outcomes in terms of mortality and length of stay. The authors considered that changes related to the clinical outcome may require a longer duration to be evident. It is of significance to study a sub-analysis of the effect of the different elements of the ASP. This, however, requires a larger sample size, and accordingly, the authors studied the impact of the program as a bundle.
CONCLUSION
Applying antibiotic stewardship programs including preauthorization of restricted antibiotics, timeouts of de-escalation and discontinuation, and implementing hospital-specific guidelines could reduce the unnecessary carbapenems use in the ICU with a subsequent decrease in the emergence of the Klebsiella resistant strains.