Impact of ventilator-associated pneumonia on mortality and epidemiological features of patients with secondary peritonitis

Intra-abdominal infections (IAIs) are one of the most important causes of mortality in the intensive care unit (ICU) [1]. Secondary peritonitis constitutes 80–90 % of cases of IAIs and is originated from the microbiological infection of the gastrointestinal tract by the perforation of hollow organs, ischemia, malignancy, and perioperative complications (anastomotic leakage, intraoperative contamination) [2, 3]. Secondary peritonitis can be classified in community-acquired and hospital-acquired, this latter associated with microorganisms presenting antibiotic resistance [2]. Mortality rate due to secondary peritonitis ranges approximately between 10 and 20 % [4–6]. During the management, the clinical outcome of the patient may be critically compromised by the development of nosocomial infections [7]. Ventilator-associated pneumonia (VAP) is a type of hospital-acquired pneumonia that is developed after at least 48 h of the patient’s intubation [8]. VAP is the most frequent of the nosocomial infections occurring in the ICU, affecting to 9–27 % of all intubated patients [9]. The VAP is associated with an increased length of hospital stay, of about 4–13 days, and hospital costs [10–13]. In our knowledge, there are only in the literature three studies analyzing specifically clinical and epidemiological aspects of the development of VAP following secondary peritonitis. The first one was a retrospective study of 1982, which reported clinical outcomes of 143 patients with intra-abdominal abscesses, and revealing an incidence of VAP of 28.7 % of the patients, and a mortality rate attributed to VAP of 65.9 % [14]. The second one was a prospective study published in 1991 comparing clinical outcomes between nosocomial pneumonia and recurrent IAI [15]. The incidence of VAP was 19.7 % of cases, and the mortality rates were 53 % for the group of patients with pneumonia and no recurrent IAI, and 75 % of those with both conditions. Finally, the third study, of 2006, included retrospectively medical records from 618,495 patients undergoing intra-abdominal surgery [16]. From them, 13,292 patients developed subsequently pneumonia, and the mortality rate was of 10.7 %.

Although there are extensive studies analyzing secondary IAIs or VAP in the ICU, in our knowledge, studies focusing specifically on VAP following secondary peritonitis are scarce and date mainly from two decades ago. Furthermore, there are some issues that remain being characterized, such as the lapse time between the starting of the peritonitis and VAP onset, and whether or not the etiologic agents responsible for IAIs are the same that for VAP, which is critical for the selection of the empirical antibiotic therapy. Early VAP onset has been associated with better prognosis, while the late one has the highest mortality rates and is often associated with multidrug-resistant microorganisms [8]. Our working hypothesis is that pneumonia increases the mortality in patients developing peritonitis. Therefore, the objective of the study was to determine in-hospital mortality and epidemiological features attributed to VAP following secondary peritonitis.

This prospective observational study involved consecutive patients admitted in the ICU of the clinical university hospital of Valladolid between May 2008 and May 2015 for the management of a secondary peritonitis. All patients, or family members, signed the written consent form to participate in the study. The collection of respiratory and blood samples, for microbiological examinations, was required for the inclusion in the study. Patients presenting primary peritonitis or those who refused to sign the consent form were excluded from the study. One of the investigators made daily rounds in the ICU to identify eligible patients and determine the onset of VAP. Because of the observational nature of the study, investigators did not interact with ICU treating physicians for the diagnosis or management of VAP. To test our working hypothesis, the primary endpoint was to evaluate whether or not VAP patients had a higher mortality rate than non-VAP patients. Secondary endpoints included the identification of variables potentially associated with in-hospital mortality and with the development of VAP. Procedures were performed in accordance with guidelines established by the hospital’s ethics committee and the Declaration of Helsinki.

Despite the significant impact of nosocomial infections on morbidity and mortality of patients staying in the ICU, no study over the last 20 years has determined the clinical, epidemiological, and microbiological features of VAP following secondary peritonitis. Therefore, the goal of the present study was to complete and update such lacking information. The most relevant results from our study included: (1) an updated incidence value of VAP of 9.6 %; (2) risk factors associated with the development of VAP including hospital-acquired peritonitis, requiring >48 h of mechanical ventilation, and SOFA score; (3) mainly late onset of VAP, and caused by multidrug-resistant microorganisms intrinsically different for each condition; (4) the 90-day in-hospital mortality rate of 47.5 % of VAP patients; and (5) independent factors associated with 30- to 90-day in-hospital mortality including VAP and SOFA score.

Overall incidence of VAP reported in our study was 9.6 % of patients who underwent surgery due to peritonitis, a value significantly lower than previous studies, ranging between 20 and 30 % [14, 15]. A possible explanation may derive from the fact that, in our study, VAP included both hospital- and community-acquired cases; however, previous studies only included hospital-acquired cases, which are associated with higher incidence rates. Moreover, the reduction in the impact of VAP over the years may be a result of the implementation of effective preventive strategies in the ICU, such as the Spanish national VAP prevention bundle called “zero VAP,” based on good general practices for control of the infection and pathogenic-tailored strategies [23, 24]. The International Nosocomial Infection Control Consortium reported a decrease in incidence from 15 to 8 % in a surveillance study conducted from 2004 to 2009 [25, 26].

There are many factors potentially associated with the development of VAP, including preexisting medical conditions in the patient (such as immunosuppression or chronic obstructive lung disease), body position, level of consciousness, nasotracheal intubation, duration of the mechanical ventilation, ventilator circuit-related factors, enteral nutrition, or personnel-related factors (such as inadequate hand hygiene or change in gloves between patients) [12, 27, 28]. In our study, hospital-acquired peritonitis, requiring >48 h of mechanical ventilation, and SOFA score were independent factors for VAP development. The duration of ventilation has been positively correlated with the development of VAP, although this potential risk seems not to be constant over the time [12]. Specifically, the risk of VAP has been estimated in 3 % per day during the first week of mechanical ventilation, 2 % in the second week, and 1 % in the subsequent weeks [27]. A high SOFA score at admission in the ICU has been associated with mortality in VAP patients [29]; for this reason, a high score in multiple organ dysfunction, concomitant with a probably immunosuppression status and/or other underlying medical conditions, may be a cause for the development of such opportunistic infection. Similarly, hospital-acquired peritonitis is associated with microorganisms presenting antibiotic resistance, poor outcomes, and longer stays in the hospital, compared with community-acquired peritonitis, which may explain its correlation with the development of VAP [2].

The onset of VAP was late in the majority of patients. More than 90 % of cases of VAP occur within the first 10 days of the intubation [30]. VAP was also developed about 16.8 days after the initiation of the peritonitis. A high percentage of the microorganisms responsible of VAP were multidrug resistant (62.5 %). Regarding the antibiogram, almost half of patients (55.0 %) received an adequate treatment. The low number of patients receiving adequate antibiotic treatment may be a consequence of: (1) the late onset of pneumonia (16.8 ± 15.1 days) since the initiation of the peritonitis. At this time, patients had received other antibiotic treatments for the peritonitis. (2) the peritonitis was the primary target of the antibiotic treatment; VAP was not so. It has been demonstrated that patients who receive inadequate empiric antibiotic treatment have longer hospital stays, higher rates of abscesses, and mortality [31]. For this reason, attention should be given to the antibiogram of each respective center, selecting adequate antibiotics taken into account potential multidrug-resistant microorganisms [1]. The high rate of multidrug resistance found in our study may be correlated with the higher proportion of patients presenting late-onset VAP. In general, etiological microorganisms responsible for the peritonitis were different than for VAP, which is consistent with previous studies [6].

According to published studies, mortality rates attributable to VAP range between 53 and 75 % [14, 15]. Since clinical outcomes depend on the length of the stay in the ICU, one goal of our study was to investigate differences in in-hospital mortality at different endpoints (30, 30–90, and 90 days). VAP patients showed a significant higher mortality rate and longer hospital stay than non-VAP. The 90-day mortality was 47.5 % of VAP patients, a value slightly lower than previous studies. Similar to the incidence, there is a decreasing tendency in the mortality rate as a result of the implementation of preventive strategies [22, 25]. In our study, the mortality rate was actually quite high, even in the non-VAP cohort, for secondary peritonitis. The recent multicenter STOP-IT trial has reported a mortality rate of 1.2 % for patients with complicated intra-abdominal infection [32]. Although this study cohort did have a much higher rate of septic shock and severe sepsis than that our study, most recent severe sepsis and septic shock studies have reported mortality rates of approximately 25 %. Risk factors associated with in-hospital mortality included VAP and SOFA score, although the univariate analysis revealed both VAP and stay at ICU as significant risk factors associated with mortality. VAP was only significant for 30- to 90-day in-hospital mortality. According to the literature, the development of VAP is associated with the percentage of nonsurvivors and survivors in a rate 2:1. It is interesting to note that Kaplan–Meier curves showed a significant divergence in survival likelihood since approximately the day 40th after admission in the ICU. Among the extensive studies evaluating the risk factors associated with worse outcomes and mortality for secondary peritonitis [31, 33–35], none of them have included VAP in their analyses. One possible reason to omit it might derive from the fact that VAP is intrinsically associated with the stay in the ICU, whatever the underlying condition of the patient. Nevertheless, results of our study highlight the importance of including VAP as a factor involved in the 30- to 90-day in-hospital mortality of patient with secondary peritonitis. Our present study had some limitations. One of them was that the study was performed in a single center. A multicenter study might have strengthen results obtained in the study and have reduced factors intrinsically associated with the center, such as the empiric antibiotic treatment or the spectrum of nosocomial pathogenic microorganisms. Another limitation of the study was the low number of patients developing VAP (40). Although similar to the sample size of the literature, a large cohort of patients might also have strengthen our results and diminish inter-individual differences.

In light of the impact on morbidity and mortality in the ICU, more attention should be given to the concurrent features among VAP and secondary peritonitis. Additional prospective studies, involving large cohort of patients, are required to corroborate these results.