Ventilator-associated pneumonia (VAP) is one of the most common Intensive Care Unit (ICU)-associated infections in patients who receive ventilation support, which would prolong ICU and/or hospital lengths of stay, and induce high mortality rate and heavy financial burden on health care services.[1] The patient outcome could considerably develop by early and precise diagnosis, more selective anti-microbial use and better routes of administration.[2,3] Piperacillin–tazobactam is a broad-spectrum β-lactam-β-lactamase inhibitor antibiotic used for the treatment of critically ill patients with VAP.[4] It is recognized that β-lactams are time-dependent antibiotics and that their effectiveness is in association with duration of free drug concentrations over the minimum inhibitory concentration (t > MIC) of organisms.[5] Although the routine mode of piperacillin administration is intermittent infusion (II), continuous infusion (CI) may also be advocated for improving the time above the MIC.[6] Animal studies have confirmed the greater efficacy of β-lactam CI over II.[5,7] However, the clinical benefit of CI or II of β-lactam antibiotics is in doubt in humans. Clinical outcomes such as mortality,[8–16] time to normalization of leukocytosis or pyrexia,[10,14,17,18] adverse events,[19] microbiological outcome,[18,20] duration of mechanical ventilation[11,15,17,20] and length of stay are comparable between bolus and continuous dosing of β-lactam antibiotics in seriously ill patients.[11,15,17,19] CI of β-lactam antibiotics may not also lead to a significant progression in the clinical cure compared with an intermittent bolus.[9–11,13,15,17,21,22] On the other hand, clinical cure[20,21] and 14-day mortality[23] of VAP patients could be significantly improved in the CI group compared with the intermittent bolus. Thus, further studies are required to evaluate the outcomes between continuous and intermittent administration of β-lactam antibiotics in critically ill patients. Here, in this study, we attempted to compare the clinical outcomes of continuous and intermittent administration of piperacillin–tazobactam by serial measurements of the Clinical Pulmonary Infection Score (CPIS).

Dosing for infections has been revised since developments of pharmacodynamic characteristics of antimicrobials and different patterns of their bactericidal activity.[26] Although the standard mode of administration of piperacillin is II, CI is of particular importance for optimizing the time above the MIC (t > MIC) in clinical cure improvement.[6] In addition, several studies have evaluated clinical outcomes including severity of illness, duration of mechanical ventilation, mortality, clinical cure from actual infection, time to normalization of leukocytosis or pyrexia and lengths of ICU stay.[8–22] Similar to other studies,[11,20,23,27] no significant differences were found between the CI (n = 31) and the II (n = 30) groups in terms of sex, age, APACHE II score at ICU admission, diagnosis and microorganism responsible for VAP in this study. Consequently, mortality is not mainly dissimilar in the continuous and intermittent dosed groups.[8–11,13,14,20,27] Rapid improvement in PaO₂/FiO₂ has been granted as the most precise marker for adequate treatment. Duration of treatment was directly associated with the CPIS score at the time of pulmonary infection diagnosis in the Micek et al. study.[28] Also, duration of piperacillin–tazobactam therapy was correlated with the CPIS score at the onset of VAP symptoms (r = 0.364, P = 0.004), similar to our study. In this regard, CPIS scores from Day 1 to Day 8 and, more importantly, from Day 1 to Day 3 were not significantly different between the groups in our study. Although we observed that the temperature, leukocyte count (on Day 8), secretions and PaO₂/FiO₂ ratio (on Day 3) improved, these changes were mostly slight. We did not find any improvement in infiltrates. Similar studies have indicated limited value of chest radiography for clinical outcome assessment in patients with pneumonia,[29–31] while quickly deteriorating pathologies are indicative of either progression or recurrence of VAP. The PaO₂/FiO₂ ratio was the only parameter that increased slightly within 3 days of the onset, especially in group II, but did not achieve a normal value. Similar to our findings, Dennesen and coworkers demonstrated that temperature, PaO₂/FiO₂ ratio and leukocyte counts improved in time after initiation of antibiotic treatment, and the resolution of these parameters was generally slow.[32] Another study used CPIS to define whether a patient was responding to therapy, with most classic parameters of infection, such as amount and quality of secretions, radiographic infiltrate, leukocytosis and fever being poor indicators of t to therapy, while a more specific physiologic marker, the PaO₂/FiO₂ ratio, was more precise.[33] Temperature rise and leukocyte count are considered as reliable criteria for supporting VAP diagnosis; however, this may lead to incorrect decision because they are nonspecific markers in severely ill patients who suffer from sepsis, shock, physical stress or acute respiratory distress syndrome, or receive medications such as corticosteroids or b-agonists.[34,35] It emphasizes the effects of confounding factors such as underlying disease, prior antibiotic therapy, other medications in combination with piperacillin–tazobactam and previous treatment in each groups on outcome. In addition, lack of established criteria in our ICU compared with other studies produced a lower CPIS score on Day 3. Moreover, ample referral patients with preoccurred VAP episodes could have skewed our estimation for timing of VAP onset. Consequently, false and late diagnosis of VAP may have misleaded our CPIS scores on Days 1, 3 and 8. In previous VAP clinical trials, a wide range of mortality rates have been reported. By reviewing VAP studies since 1987, the rough mortality rates ranged from 24% to 76%. This wide range probably reflects differences in patient characteristics, underlying disease, diagnostic criteria and the pathogens involved.[36] According to another study, the mortality of Pseudomonas or Acinetobacter species-associated pneumonia was 87%, compared with a 55% mortality rate of pneumonia due to other organisms.[37] Similarly, Kollef et al. reported that patients with high-risk pathogens causing VAP (Pseudomonas aeruginosa, Acinetobacter spp and Stenotrophomonas maltophilia) had a considerably higher hospital mortality rate (65%) than VAP patients with other organisms (31%).[38] The mortality of 60.7% in the whole population, 66.7% in the II group and 54.8% in the CI group, of our study was close to those reports. Pseudomonas aeruginosa and Acinetobacter sp were detected as the most common MDR pathogens in our study. In this study, the etiology of VAP isolated from the BAL specimen during the course of piperacillin–tazobactam therapy on Days 1, 3 and 8 did not significantly vary between the two groups. The predominant Gram-negative bacteria were Acinetobacter sp and Pseudomonas aeruginosa, followed by Klebsiella sp, Escherichia coli and Enterobacter sp. Several studies have reported that more than 60% of VAP is caused by aerobic Gram-negative bacteria.[39,40] The limitations of our study include the fact that the two modes of administration were not compared using a randomized design, piperacillin–tazobactam serum concentrations were not determined and the microbiological and laboratory data were not available as soon as other clinical information or at the favorite days. Also, it was performed within a single ICU and in a small sample.

In conclusion, there were no significant differences in the clinical outcomes of patients receiving piperacilin/tazobactam via CI or II by serial measurements of CPIS score. Therefore, we should emphasize that before choosing the standard mode of administering β-lactams, more randomized clinical trials are necessary to establish the potential advantage of administering β-lactams by CI or II in accordance with the MIC of the microorganism responsible, the ratio of tissue and serum concentration, the volume of distribution of the antimicrobial agent and the stability of the antimicrobial agent once dissolved between the two groups.