Owing to critically ill patients’ unstable hemodynamic status, antimicrobial therapy for the management of severe infections is challenging. Antimicrobial selection, in particular, proves difficult as every delay increases patients’ mortality risk. In critical care settings, antimicrobial selection involves numerous problems, the most significant among which are suitability in terms of disease pathophysiology and the need to consider patient characteristics, such as renal and hepatic function, hemodynamic status, and risk of adverse drug reactions. Therefore, inappropriate antimicrobial use increases the chances of adverse drug reactions, Clostridioides difficile infections, and multidrug-resistant pathogen development (Brusselaers et al., 2011; Taggart et al., 2015). In this context, antimicrobial stewardship strategies are the only evidence-based solution as per the recommendations of global health regulatory agencies.

A systematic review by Kaki et al. (2011) demonstrated that antimicrobial stewardship programs (ASPs) in intensive care units (ICUs) can be successful through the use of institution-specific strategies. The majority of these studies focused on the impact of ASPs on antimicrobial consumption and cost. However, for ASPs in critical care settings to be effective, in addition to ensuring adherence to institutional goals and strategies, they must promote prompt and rational antibiotic use for life-threatening infections. An ICU ASP may exhibit some very ICU-specific goals and strategies (Brusselaers et al., 2011; Kaki et al., 2011). According to the Centers for disease Control and Prevention (CDC), ASPs in healthcare institutions must be multidisciplinary, encompassing involvement of leadership, clinical pharmacy, microbiology, and clinician education and training regarding infectious disease state management, recent antimicrobial use and trends, and optimal prescribing strategies (Pollack and Srinivasan, 2014).

As there is little margin for error with regard to ICU patients, the early initiation of broad-spectrum antimicrobial therapy to cover all possible pathogens and subsequent targeting of infective pathogens based on clinical and microbiological findings has been indicated (Prowle et al., 2011). Based on this concept, according to the National Health Services strategy (Start Smart and Then Focus) for the implementation of successful ASPs in secondary care, it is important to emphasize the initiation of effective antimicrobial therapy in the first hour of admission among critically ill patients and review the decision in 48 h. Antimicrobial therapy is then either to be continued or stopped through intravenous (IV) to oral conversion, de-escalation to narrow-spectrum agents, or switching to outpatient parenteral antibiotic therapy (Ashiru-Oredope et al., 2012, Supplementary Material). In addition, clinical decision support systems are vital in acute care settings to ensure appropriate and timely antimicrobial delivery for life-threatening infectious conditions (Lawrence and Kollef, 2009).

In Saudi Arabia, the concept of antimicrobial stewardship has been gaining ground since the 2014 introduction of the implementation process by the Ministry of Health’s General Administration of Pharmaceutical Care at public sector hospitals. According to their strategic plan, as of 2017, the ASP implementation process was in stage 4. One of the barriers to successful ASP implementation in the Saudi context is the lack of human resources (Alomi, 2017). To date, there have been limited Saudi studies evaluating the effectiveness of multidisciplinary ASPs in critical care settings (Amer et al., 2013; AlAwdah et al., 2015). Therefore, we recently introduced a multidisciplinary ASP in a critical care setting by adopting the CDC’s core antimicrobial stewardship elements and the National Health Services’ Start Smart and then Focus strategy. The national ASP toolkit was modified according to the hospital population and local resistance pattern in Makkah city. A prominent factor in the program was the involvement of higher administration at the leadership level and the use of education as a supplementary strategy to utilize current human resources. We used quasi experimental study to evaluate the impact of multidisciplinary antimicrobial stewardship interventions on antimicrobial use with cost and clinical outcomes as the primary and secondary outcome, respectively.

During the pre-intervention phase, 684 patients were admitted to the ICU, among whom 135 (19.7%) had confirmed infectious diseases. Similarly, in the post-intervention period, there were 623 patients admitted to the ICU, among whom 169 (27.1%) had confirmed infectious diseases. There were no significant differences in gender, age, nationality, and final diagnosis between the pre- and post-intervention periods (Table 1). There were few differences in comorbidities between the two groups.

The mean total monthly antimicrobial utilization decreased by 25% from 742.86 DDD per 100 bed days to 555.33 DDD per 100 bed days after ASP implementation (p = 0.110). There was a significant decrease in total IV antimicrobial usage by 129 DDD per 100 bed days after the intervention (p = 0.038). There was no significant change (p = 0.347) in oral antimicrobial usage in the post-intervention period. With respect to specific antimicrobials, there was a significant drop in ceftriaxone use; there was a mean reduction of 77.64 (p = 0.008) DDD per 100 bed days. The mean total monthly consumption of oral levofloxacin was significantly increased by 145.54 DDD per 100 bed days from 26.75 DDD per 100 bed days to 172.29 DDD per 100 bed days (p = 0.008). There were no significant changes (p = 0.767) in the use of H2 blockers or proton pump inhibitors in the ICU between the two study phases (Table 2).

There was no significant change in cost as Saudi Riyals (US $1.00 = 3.74 SR) between the periods. The mean total cost of antimicrobials increased from 38,723.72 Saudi riyals per 100 bed days before the intervention to 87,552.70 Saudi Riyals per 100 bed days after the intervention (p = 0.678). This non-significant change was due to the inclusion of linezolid in the formulary in the post-intervention phase. There was a significant decline in total mean cost (37%; p = 0.008) when the cost of linezolid was not considered in the post-intervention phase. There was a significant change in the mean cost of IV and oral antimicrobials, with a mean reduction of 49,486.77 Saudi Riyals (p = 0.021) for IV antimicrobials and 657.8 Saudi Riyals (p = 0.021) for oral antimicrobials after ASP implementation (Table 3).

There were no significant changes in infectious disease-related mortality rate, length of stay in the ICU, and proportion of patients admitted and discharged from ICU between the pre- and post-intervention periods (Table 4). The most common ASP recommendation was IV to oral conversion, addition of antibiotics, and de-escalation. Almost all interventions were accepted by physicians, with a few disagreements according to patients’ clinical conditions (Table 5).

This study marked the adoption of CDC core elements to facilitate the implementation of an ASP in an acute care setting. The multidisciplinary approach resulted in significant outcomes because of the involvement of higher administration. An immediate drop in restricted antimicrobials was observed, with overall decreases in piperacillin + tazobactam (66.5%), ceftriaxone (82.3%), and vancomycin (66%) after ASP implementation. On the contrary, there was a substantial increase in the usage of antibiotics such as meropenem (54%), linezolid (100%), and IV levofloxacin (66%) owing to the timely delivery of appropriate antimicrobials in the context of life-threatening infectious diseases like sepsis and severe community acquired pneumonia. In the pre-intervention phase, ceftriaxone was the drug of choice for most intensivists in the ICU. Similarly, vancomycin was the only anti-methicillin-resistant Staphylococcus aureus (MRSA) agent for ICU patients. Linezolid was introduced in the formulary to cover infections with suspected MRSA origin in patients with renal instability.

No significant immediate reductions were observed for control medications, suggesting that the interventions were responsible for change in antimicrobial use. However, there was no appreciable change in total mean cost for antimicrobials in the post-intervention phase, mainly because of the formulary’s introduction of linezolid, the costliest antibiotic among all classes. In addition, meropenem and levofloxacin use also contributed positively to the rise in costs.

Overall, the magnitude of variation in antibiotic usage was similar to the findings reported in a systemic review in critical care settings (Kaki et al., 2011). Similar changes in total antimicrobial use were observed in a Hungarian ICU with restriction policy as the major intervention. Antibiotic consumption decreased by 37.8% (from 162.9 to 101.3 DDD per 100 patients days) (Peto et al., 2008). Similarly, in a Canadian study, the implementation of an audit and feedback strategy in the ICU led to a reduction in mean total antimicrobial use in DDD per 100 bed days by 28% (Taggart et al., 2015). Likewise, in studies in France (Geissler et al., 2003) and Germany (Meyer et al., 2007), similar reductions in total antimicrobial use were observed.

As the main goal of ASP follow-up was to restrict all antimicrobials to the right patients, reductions were mainly observed for restricted antibiotics with compensatory increases in the use of non-restricted antibiotics, for example, levofloxacin for the initial management of community acquired pneumonia. This phenomenon is commonly known as the “squeezing-the-balloon” effect (Burke, 1998). Therefore, the strength of our study lies in the overall reduction of all antimicrobial use, relatively than specific antibiotics only.

As antimicrobial resistance relates to the selective pressure exerted by the overuse of antibiotics (Herwaldt and Duncan, 1997), based on pre-intervention findings, we propose changes in the use of third-generation cephalosporin to improve antibiotic susceptibility and reduce infections related to hospital mortality in critically ill patients (Du et al., 2003). In our study, one of the justifications for the reduced use of third-generation cephalosporin was the implementation of a pneumonia protocol through prescribing forms. According to this form, the first choice in most cases was a high dose of levofloxacin, a respiratory quinolone. In addition, ceftriaxone with azithromycin was not a good choice in cardiac patients because of the risk of azithromycin-induced arrhythmias (Ray et al., 2012; Albert et al., 2014).

Another important finding in our study was the 82.3% reduction in the use of ceftriaxone. This is important as irrational selection of initial antimicrobials could inevitably enhance the chance of mortality in patients with life-threatening infections (Torres et al., 1990; McGowan, 1994). Recent guidelines for the treatment of severe community acquired pneumonia generally recommend levofloxacin (Dunbar et al., 2003; Mandell et al., 2007). Therefore, based on this rationale, the pre-intervention use of ceftriaxone, and the fact that a change in empirical antibiotics may reduce resistance to common enterobacteriaceae-associated infections and improve patient management in the ICU, we focused on replacing ceftriaxone with levofloxacin for empirical therapy for community acquired pneumonia (Mebis et al., 1998). Most intensivists welcomed this strategy and accepted the majority of the ASP team’s recommendations.

Our findings were consistent with a previous result regarding the impact of ASPs on clinical outcomes. In a systematic review by MacDougall and Polk (2005), limiting antibiotic use had no significant impact on an institution’s clinical outcomes (Peto et al., 2008) (MacDougall and Polk, 2005; Peto et al., 2008). Interestingly, no cases of Clostridium difficile infection were reported in either the pre- or post-intervention period.

The total increase in mean antimicrobial use achieved through the implementation of the ASP was similar to an ICU study from other Saudi regions. In a study by Amer et al. (2013), ASP interventions caused a reduction in antibiotic use from 2403.64 DDD per 1000 patient days to 376.2 DDD per 1000 patient days. While antimicrobial consumption was lower in our study, our economic outcomes are similar to their findings (Amer et al., 2013; AlAwdah et al., 2015).

In summary, this study supports the use of CDC core elements for the successful implementation of an institutional ASP. Education and training along with the involvement of an ICU team with basic infectious disease training and support from leadership are effective tools for the achievement of antimicrobial stewardship goals.