A systematic review and meta-analysis of published studies was performed that compared the therapeutic efficacy of BL monotherapy with FQ monotherapy for PA infection in adult inpatients with the purpose of identifying apparent trends with respect to mortality, bacteriological eradication, and clinical success. The results demonstrated patients receiving FQ monotherapy had higher survival in PA bacteremia, but not higher rates of bacteriological eradication in pneumonia and skin and soft tissue infection caused by PA. Based on one study’s data of 26 total patients, FQ monotherapy was indicated to have no benefit over BL monotherapy on clinical success rates in PA pneumonia .
Unadjusted data were used in the mortality analysis due to the lack of adjusted data accounting for potential confounding variables. Therefore, other inherent risk factors for mortality were not accounted for in the reported data. Wu et al reported patient demographic data specifically for the BL and FQ arms. The BL arm had a greater percentage of patients with septic shock, immunosuppression, and higher mean Pitt bacteremia and APACHE II scores, whereas the FQ arm had a greater percentage of patients with malignancy . Differences in Pitt bacteremia and APACHE II scores were statistically significant, meaning the BL group had more critically ill patients . Unfortunately, treatment arm-specific patient demographics were not reported in the other two studies. In addition to the retrospective nature of the included studies, these discrepancies may have biased our results to an unknown extent.
For the meta-analysis on bacteriological eradication, all included studies reported the presence of polymicrobial infection, which often occurs in patients with ventilator associated pneumonia . This fact likely increased the external validity of the results. All three studies included an unknown percentage of patients who were allowed to receive non-protocol antimicrobials. Two of three studies allowed for the option for protocol-allowed vancomycin and/or metronidazole. These study designs may have introduced a certain level of risk for bias in the results. Although treatment arm-specific patient demographics were reported in all the studies, none of the studies further stratified these data for infection caused by PA specifically. Therefore, we could not determine if any confounding variable existed that would impart a potential advantage or disadvantage for bacteriological eradication.
It is worth noting the Clinical & Laboratory Standards Institute (CLSI) has changed the minimum inhibitory concentration (MIC) cutoffs for ciprofloxacin over time. Since the 1990’s, the ciprofloxacin MIC cutoff has decreased, meaning the criteria to be considered “active therapy” is different now for more contemporary studies than it used to be for the studies from the 1990’s. Now, it is unclear how this change could have affected the results, but it is important to acknowledge the CLSI change.
Strengths of this meta-analysis include adherence to a well-defined PICOT model, the use of NIH quality assessment tools, and following PRISMA reporting guidelines for systematic reviews and meta-analyses. Additionally, the variations in study year, geography, and patient demographics among included studies enhance the external validity of this study. However, limitations include small sample sizes, differing definitions of terms and outcomes among studies, incomplete data, the lack of grey literature search, and known and unknown discrepancies in patient demographics between treatment arms for infections caused specifically by PA. The latter is explained by the fact that not all studies were designed with the purpose of making direct comparisons between BL and FQ monotherapy or studying PA-specific infection (rather, some examined all causative pathogens of a given infection type). Additionally, the studies evaluated definitive therapy, and the efficacy of empiric therapy could have impacted the results. Consequently, these results should be interpreted with caution.
Overall, little research has focused specifically on comparing the therapeutic efficacy of BL and FQ drug classes as active, definitive monotherapy for PA infection. To our knowledge, this is the first meta-analysis to do so. The vast majority of systematic reviews compares combination therapy and monotherapy, but results are controversial [1, 11]. Although combination therapy is recommended in certain cases of PA infection for empirical therapy [10, 11], de-escalation to a single active agent is encouraged as this may decrease the potential for adverse events and antimicrobial-associated toxicity and reduce the development of resistance . De-escalation to monotherapy is also consistent with antibiotic stewardship program objectives . Regardless, more research comparing monotherapy in PA infection are needed.
When selecting a preferred drug class for PA infection, no overarching recommendations exist, and our results do not come close to bridging the gap. More likely, no one drug class or antimicrobial agent is the ideal choice. Antimicrobial resistance is in flux and varies by regions, so local antibiogram data are essential to selecting drug therapies. Once PA antimicrobial susceptibility testing is complete, the interplay between patient characteristics and drug features remains critically important in selecting definitive therapy. Furthermore, definitive antimicrobial therapy for PA must minimize the potential for selecting resistance.
On a whole, these results provide insight into the therapeutic efficacy of BL and FQ drug classes as active, definitive monotherapy for PA infection in adult inpatients but fall short of offering definitive answers. The results suggest FQ monotherapy is associated with significantly higher survival rates compared to BL monotherapy, but more rigorous research are required to make definitive conclusions. Clinicians should continue to weigh pros and cons of drug classes and individual agents for a particular patient when selecting definitive therapy for PA infection.