Abstract
Abstract
Background:
Patients with complicated intra-abdominal infections (cIAIs) caused by multi-drug–resistant organisms (MDROs) have been identified as being at increased risk for adverse outcomes. Prompt identification and stratification of these patients is essential in the clinical management, allowing the physician timely optimization of empiric antimicrobial therapy while awaiting results of intra-operative cultures to streamline antibiotic treatment.
Methods:
The study is a secondary analysis from two prospective multi-center color surveillance studies. It included all consecutively hospitalized adult patients undergoing surgical procedures, interventional drainage, or conservative treatment with cIAIs, with positive cultures performer on intra-operative samples of peritoneal fluid or purulent exudate/discrete abscesses. Patients with pancreatitis and primary peritonitis were excluded. A case-control approach has been used to evaluate the factors associated with the isolation of a MDRO in enrolled patients.
Results:
Among 1986 patients included in the study, a total of 3534 micro-organisms were isolated from intra-peritoneal fluid samples; in 46.5% of cultures, two or more pathogens were identified. The MDROs represented 9.8% of the total of isolated micro-organisms. The overall incidence rate of MDROs was 13.9%. The MDROs were more frequently isolated in patients with health-care–associated cIAIs (25.4%). Multi-nomial logistic regression analysis of risk factors demonstrated that statistically significant risk factors independently associated with the occurrence of MDROs were previous antimicrobial therapy administered within seven days before operation, presence of severe cardiovascular disease, white blood cell count <4000/mL or >12,000/mL, cIAI acquired in a healthcare setting, and inadequate source control.
Conclusions:
The study showed that knowledge of five easily recognizable variables—assessable on hospital admission or as soon as the surgical intervention is concluded—might guide the surgeon to identify patients with cIAIs caused by MDROs, and therefore to choose the most adequate empiric antimicrobial therapy for them.
I
An increasing proportion of multi-drug–resistant organisms (MDROs) has been reported in IAIs [6,11,12]. It is critical to take into account factors that influence abdominal bacterial ecology and the susceptibility of causative organisms to ensure optimal management. The risk factors most consistently associated with MDRO colonization or infection are: Age greater than 70 years; immunosuppression; chronic dialysis; numerous comorbidities; admission from a healthcare facility; recent hospitalization; a recent use of antibiotic agents, including beta-lactams and quinolones; recent surgical procedure; recent urinary catheterization; and history of MDRO colonization [13–17].
To assess the microbiologic differences, particularly with respect to the type of bacteria recovered and the level of antimicrobial susceptibility between community-acquired IAIs (CA-IAIs) and health-care–acquired IAIs (HA-IAIs), several epidemiologic studies have been published, at an inernational level [18–20], or limited to a single hospital, region, or country [3,7,11,12,21,22]. Only a few observational studies were published, however, with the purpose to investigate patient characteristics associated with a high risk of isolation of resistant pathogens from an intra-abdominal source [2,12,21,23].
The objectives of this study were: (1) To analyze a large-scale multi-center dataset of cIAIs to assess the microbiologic differences between community-acquired complicated intra-abdominal infections (CA-cIAIs) and health-care–acquired complicated intra-abdominal infections (HA-cIAIs), particularly with respect to the type of micro-organism recovered and the patterns of antimicrobial susceptibility; (2) to identify specific risk factors for MDROs isolation in patients with cIAIs, which could help to improve the adequacy of empiric antimicrobial therapy.
Methods
Study population and design
This is a secondary analysis from two prospective multi-center cohort surveillance studies: The Complicated Intra-Abdominal infection Observational (CIAO) study [24], and the Complicated Intra-Abdominal infection Observational Worldwide (CIAOW) study [25], performed between January 2012 and June 2012, and between October 2012 and March 2013, respectively. These studies included all consecutively hospitalized adult patients (older than 18 years) undergoing surgical procedures, interventional drainage, or conservative treatment with cIAIs (defined as abdominal infections originating in an organ cavity, extending into the peritoneal space, and forming an abscess or peritonitis [1]), with positive cultures performer on intra-operative samples of peritoneal fluid or purulent exudate/discrete abscesses. Patients with pancreatitis and primary peritonitis were excluded.
The CIAO study was performed in 68 medical institutions throughout Europe, and included 2152 patients, whereas the CIAOW study was performed in 68 medical facilities worldwide, and enrolled 1898 patients. The purpose of these studies was to describe the clinical, microbiologic, and treatment profiles of both CA-cIAIs and HA-cIAIs in a European context and in a worldwide context. In the present study, a case-control approach has been used to evaluate the factors associated with the isolation of a MDRO in enrolled patients.
Data collection
On admission, the following data were collected from each patient: Gender, age, antimicrobial therapy administered within seven days before operation, presence of comorbidities (primary or secondary immunodeficiency, solid or hematopoietic and lymphoid malignancy, severe cardiovascular disease). Clinical findings, such as fever (defined as core temperature >38.0°C) or hypothermia (core temperature <36.0°C), leucocytosis (white blood cell count [WBC] >12,000/mL) or leukopenia (WBC <4000/mL), and the presence of sepsis, severe sepsis, or septic shock were recorded. Complicated IAIs were classified as CA-cIAIs or HA-cIAIs. The HA-cIAIs have been diagnosed in patients hospitalized for at least 48 hours during the previous 90 days; those residing in a nursing or long-term–care facility during the previous 30 days; those who have received intravenous therapy, wound care, or renal replacement therapy within the preceding 30 days. Moreover, HA-cIAIs were classified as chronic care setting acquired, post-operative, and non–post-operative hospital setting acquired. The source of infection (stomach or duodenum, cholecyst, small bowel, colon, appendix, or other) and peritonitis diffusion (generalized or localized peritonitis/abscess) were determined.
Source control (conservative treatment, operative or non-operative interventional procedures) and its adequacy were noted, defining the latter as achieving establishment of the cause of cIAIs and controlling the origin of peritonitis [9]. The delay in the initial intervention was established if the time elapsed between admission and the performance of source control was greater than 24 hours. Duration of empiric antimicrobial therapy and its adequacy were also recorded, considering the latter as the patient's clinical and laboratory improvement after the beginning of therapy according to the discretion of the providers for each case. Re-operation during the hospital stay, the length of intensive care unit (ICU) and hospital stay, and hospital deaths were registered.
Microbiologic results from cultures performed on intra-operative samples of peritoneal fluid or purulent exudate/discrete abscesses to identify gram-negative, gram-positive, and anaerobes bacteria and fungi were collected. Every hospital center determined antimicrobial susceptibilities of the isolated micro-organisms according to its own procedures and criteria. According to the joint recommendations for epidemiologic studies from the European Centre for Disease Prevention and Control, and from the Centers for Disease Control and Prevention [26], every isolated micro-organism showing non-susceptibility to at least one agent in three or more antimicrobial categories of antimicrobial agents was classified as a MDRO. In our study, MDROs were classified as follows: Acinetobacter baumanii resistant to carbapenems, Escherichia coli producing an extended-spectrum beta-lactamase (ESBLp), Klebsiella oxytoca ESBLp, K. pneumoniae ESBLp or resistant to carbapenems, Pseudomons aeruginosa resistant to carbapenems, Enterococcus faecalis resistant to glycopeptides, E. faecium resistant to glycopeptides, methicillin-resistant Staphylococcus aureus (MRSA), Bacteroides spp. Resistant to metronidazole, Clostridium spp. Resistant to metronidazole, Candida albicans resistant to fluconazole, other Candida spp. Resistant to fluconazole.
Statistical analysis
Data were analyzed in absolute frequency and percentage, in the case of qualitative variables. Quantitative variables were analyzed as medians and interquartile ranges (IQR). The characteristics of patients with a diagnosis of cIAI caused by a MDRO, and those with a diagnosis of cIAI caused by a non-MDRO, were compared using the Wilcoxon rank sum test for continuous variables, and the chi-square test for categoric variables.
To identify risk factors independently associated with the occurrence of MDROs, a multi-nomial logistic regression analysis was performed considering only independent variables that were able to be considered by the physician before starting an empiric antimicrobial therapy and before obtaining microbiologic results. We selected only independent variables that had a p value <0.05 in the univariable analysis. Then, a backward selection method was applied to select a limited number of variables, using a likelihood-ratio test for comparing the nested models (α = 0.05). At each step, we removed from the previous model the variable with the highest p value greater than α, checking the fit of the obtained model, and then stopping when all p values were less than α.
It was regarded as statistically significant when the p value was <0.05. All statistical calculations were performed using Stata 9 software package (StataCorp, College Station, TX).
Results
During the study period, a total of 4040 cIAIs were recorded; 1511 (37.4%) patients were treated without performing a culture of the peritoneal fluid; among them, 1391 (1391/1511, 92.1%) had a CA-cIAI versus 120 (120/1511, 7.9%) who had a HA-cIAI.
Overall, 2529 (62.6%) patients had intra-operative microbiologic cultures; 1954 (77.3%) were obtained from patients presenting with CA-cIAIs, whereas 575 (22.7%) were from patients with a diagnosis of HA-cIAIs. Among them, 543 (543/2529, 21.5%) had cultures without growth, and 1986 (1986/2529, 78.5%) had positive cultures obtained from intra-operative samples of peritoneal fluid or purulent exudate/discrete abscesses.
Microbiologic differences between CA-cIAIs and HA-cIAIs
Among 1986 patients included in the study, a total of 3534 micro-organisms were isolated from intra-peritoneal fluid samples. In 1062 (53.5%) cultures, only one micro-organism was isolated, while in 924 (46.5%) cultures, two or more pathogens were identified. The median number of micro-organisms recovered per patient was 1 (IQR 1–2).
The most frequent micro-organism isolated in peritoneal fluid was E. coli, followed by E. faecalis and K. pneumoniae, with differences observed according to the source of infection (Table 1).
GD = gastroduodenal; CNDP = colonic non-diverticular perforation.
Three-hundred forty-seven (347/3534, 9.8%) MDROs were isolated; 161 (46.4%) MDROs were isolated in patients with a diagnosis of CA-cIAIs, while 186 (53.6%) were in patients with a diagnosis of HA-cIAIs. The proportion of MDROs in the total number of isolated micro-organisms of the same species was higher in patients with HA-cIAIs compared with patients with CA-cIAIs. The most frequently isolated micro-organism was E. coli ESBLp, followed by K. pneumoniae ESBLp. Resistance phenotypes of 347 MDROs isolated in cIAIs, grouped according to the setting of acquisition, are reported in Table 2. Differences were observed according to the geographic origin of patients. The risk of isolation of MDROs was higher in patients with cIAI from the Eastern Mediterranean Region (21.5%, p < 0.05) and South-East Asia Region (28.8%, p < 0.001) compared with patients from other World Health Organization (WHO) Regions. Similarly, the risk of isolation of ESBLp Enterobacteriaceae was higher in patients with cIAI from the Eastern Mediterranean Region (19.3%, p < 0.05) and South-East Asia Region (27.3%, p < 0.001) compared with patients from other WHO Regions.
MDROs = multi-drug–resistant organisms; CA-cIAIs = community-acquired complicated intra-abdominal infections; HA-cIAIs = healthcare-acquired complicated intra-abdominal infections; ESBLp = extended-spectrum beta-lactamase producer.
Variables associated with MDROs isolation
Among 1986 cIAIs, 1434 (72.2%) were classified as CA-cIAIs, whereas 552 (27.8%) were HA-cIAIs. The most frequent source of infection in CA-cIAIs was complicated appendicitis, followed by cholecystitis and gastroduodenal perforation, while in HA-cIAIs it was post-operative, followed by small bowel perforation and other sources of infections, as reported in Table 3.
CA-cIAIs = community-acquired complicated intra-abdominal infections; HA-cIAIs: health-care–acquired complicated intra-abdominal infections; CS = chronic-setting–acquired, POHS = post-operative hospital-setting–acquired; NPOHS non–post-operative hospital-setting–acquired; PID = pelvic inflammatory disease.
Patient characteristics, detected or assessable before starting an empiric antimicrobial therapy, and grouped according to the presence or absence of MDROs, are reported in Table 4. The overall incidence rate of MDROs in 1986 patients with cIAIs was 13.9% (276/1986). The MDROs were isolated more frequently in patients with HA-cIAIs (140/552, 25.4%) than CA-cIAIs (136/1434, 9.5%). To analyze the association between patients' variables and cIAIs caused by MDROs, univariable analysis was performed.
MDRO-cIAI: complicated intra-abdominal infection caused by multi-drug–resistant organisms; IQR = inter-quartile range; HA-cIAI = health-care–acquired complicated intra-abdominal infection; PID = pelvic inflammatory disease.
All data are expressed in n (%) unless otherwise noted. All p values were calculated using two-sided chi-square test unless otherwise noted.
Wilcoxon rank sum test.
Inadequate empiric antimicrobial therapy was more frequently associated with cIAIs caused by MDROs, as well as a longer duration of antimicrobial therapy. Patients who reported a recent antibiotic exposure within the previous seven days before operation had a greater likelihood to receive an inappropriate empiric antimicrobial therapy (114/546, 20.9% vs. 158/1282, 12.3%; p < 0.001) during the hospitalization compared with patients without recent antibiotic exposure. Moreover, a statistically significant association was observed between a duration of antimicrobial therapy longer than 10 days during the hospitalization and an inadequate source control (117/183, 63.9% vs. 645/1699, 38.0% in patients with adequate source control, p < 0.001). Patients with MDROs were more likely to be admitted to an ICU and had a longer length of hospital stay. The hospital death rate was higher in patients with a cIAI caused by a MDRO, as reported in Table 5.
MDRO-cIAI: complicated intra-abdominal infection caused by multi-drug–resistant organism; IQR = inter-quartile range; HA-cIAI = health-care–acquired complicated intra-abdominal infection; ICU = intensive care unit.
All data are expressed in n (%) unless otherwise noted. All p values were calculated using two-sided chi-square test unless otherwise noted.
Wilcoxon rank sum test.
Results of multi-nomial logistic regression analysis of risk factors for cIAIs caused by MDROs are reported in Table 6. They demonstrated that statistically significant risk factors independently associated with the occurrence of MDROs were previous antimicrobial therapy administered within seven days before operation, presence of severe cardiovascular disease, WBC <4000/mL or >12,000/mL, cIAI acquired in a healthcare setting, and inadequate source control.
OR = odds ratio; CI = confidence interval; HA-cIAI = health-care–associated complicated intra-abdominal infection.
Analyzing the risk of isolation of MDROs, a statistically significant difference was observed between patients with at least one risk factor (15.4%), and patients with not one risk factor (6.1%, p < 0.001). Further, we observed a progressive increase in the risk of isolation of MDROs as the number of risk factors augmented, becoming 22.7% for patients with two risk factors, 32.2% for patients with three risk factors, 38.6% for patients with four risk factors, and 50.0% for patients with five risk factors. A statistically significant difference within all the groups was observed (p < 0.001).
Discussion
In this multi-center study, we found that up to 13.9% of patients in whom a cIAI developed had a MDRO isolated, with a significant difference observed between health-care–associated and community-acquired infections, being more frequent as expected in the former group. This is a concerning global problem that requires urgent attention. Our study shows that cIAIs caused by MDROs were significantly associated with longer duration of antimicrobial therapy and hospital length of stay, higher access to ICU and re-laparotomy. Previous studies have demonstrated that all of these factors have been associated with higher hospitalization costs and, in some cases, lower quality of care.
Empiric antimicrobial therapy in cIAIs should be based on several factors that every surgeon must take into account: Local epidemiology, individual patient risk factors for difficult to treat pathogens, clinical severity of infection, and infection source [28]. Although the empiric antimicrobial therapy should be started as soon as possible in patients with cIAIs, the prescriber always has to consider that an ineffective or inadequate antimicrobial regimen is one of the variables more strongly associated with adverse outcomes in critically ill patients. Further, selection of empiric antibiotic therapy should take into account the composition of the microbial flora, which varies according to the source of infection, with a prevalence of gram-positive bacteria and Candida spp. In the high gastrointestinal tract and a progressive increase of anaerobes and gram-negative bacteria toward the lower gastrointestinal tracts [22].
Our study showed that the etiologic distribution of micro-organisms present in the peritoneal fluid varies according to the source of infection. Almost half of the cultures were polymicrobial, and gram-negative bacteria were the most frequent isolated micro-organisms in all the samples of peritoneal fluid, of which one-third were E. coli.
In this study, we observed that risk of isolating a MDRO was independently associated with previous antimicrobial therapy administered within seven days before operation, presence of severe cardiovascular disease, leucocytosis or leukopenia, a health-care–acquired setting of infection, and an inadequate source control of the infection.
An antimicrobial therapy administered within seven days before operation was found to be a risk factor for MDROs in patients with cIAIs, with a statistically significant association in multi-variable analysis. Previous studies have demonstrated an association between prior antibiotic exposure and subsequent infection with MDROs. Seguin et al. [12], in a prospective observational study performed on 93 adult patients with a diagnosis of secondary peritonitis, observed in multi-variable analysis that a composite variable associating pre-operative hospital length of stay and antimicrobial therapy administered for two days in the 15-day period preceding secondary peritonitis was a significant independent risk factor for infection with MDROs.
In another prospective observational study performed on 115 adult patients with diagnosis of post-operative peritonitis necessitating intensive care [21], the authors concluded that only antimicrobial treatment in the three months preceding hospitalization and duration between first operation and re-laparotomy were independent risk factors for isolation of MDROs. Augustin et al. [23], in a perspective observational study performed on 100 patients with post-operative peritonitis admitted to the ICU detected that the use of broad-spectrum antibiotic agents between initial intervention and re-operation was the only significant risk factor for emergence of MDROs.
A recent exposure to antimicrobial agents must be considered as a risk factor for resistance, although a cut-off threshold having the best sensitivity and specificity for predicting the acquisition of MDROs has not been determined. In the published medical literature, various timelines have been suggested about what defines “recent” antibiotic therapy, and within the previous 90 days is a consistently cited number [28]. Patients' antimicrobial history is of key importance: Previous antimicrobial therapy strongly modified the intra-abdominal microbiota [29] and is associated with an increase in the risk of infection with a micro-organism resistant to it [30–32]. Surgeons should always identify the presence of previous antibiotic exposure as an important consideration when prescribing empiric antibiotic therapy to patients with cIAIs. Indeed, we observed also a statistically significant association between patients with recent antibiotic exposure and inadequate empiric antimicrobial therapy, which has been shown to be associated with a longer length of hospital stay and with increased clinical failures and higher death rates [6,33,34].
Rapid microbiologic tests provide opportunities for antimicrobial stewardship programs to improve antimicrobial use and clinical and economic outcomes [35]. Prompt detection of MDROs carriage based on screening methods providing a same-day turnaround time could be used for these purposes in selected patients with a diagnosis of cIAI. A possible approach might be the screening of patients reporting or showing high risk factors for infections caused by MDROs, such as an antimicrobial therapy administered within seven days before operation, on hospital admission. This strategy might be useful not only to implement additional infection control measures quickly and minimize cross-transmission, but also to enable timely antimicrobial optimization, which, in turn, may lead to decreased deaths, shortened hospital stay, and lower hospitalization costs. Further studies should be performed to help determine how best these rapid tests could be implemented for these purposes in surgical settings.
Moreover, in the present study, an inadequate empiric antimicrobial therapy was significantly associated more frequently with cIAIs caused by MDROs. The clinical challenge for the surgeon is to find the balance between ensuring that each patient is covered appropriately for the most likely pathogens of their IAI, while avoiding the use of overtly broad-spectrum antimicrobials to preserve them for future use. Antimicrobial therapy is paramount in the management of cIAIs, especially in patients with severe sepsis with or without septic shock who require both immediate and effective empiric antibiotic therapy [36]. Once the microbiologic results are available, the initial therapy has to be adjusted following the antibiotic susceptibility of the identified pathogen.
Timing and adequacy of source control are the most important issues in the management of IAIs, because inadequate and late control of the septic source—which can be achieved either by non-operative or operative interventional procedures—has been associated with increased deaths in patients with cIAIs [36,37]. Both inadequacy and delay of source control have proved to be independent risk factors for cIAI caused by a MDRO. When the infectious process induces tissue necrosis, or when necrotic tissue becomes secondarily infected, the inaccurate excision of the necrotic tissue neither removes the bulk of the bacterial burden nor leaves vascularized tissue to which antibiotic agents can be delivered [38]. In this situation, a sub-optimal target site concentration may have important clinical implications, in particular for micro-organisms for which in vitro minimum inhibitory concentrations are high [39], allowing resistant pathogens to propagate and cause the infection.
Our study has some limitations. We conducted an observational study involving a large, but probably not representative, number of hospitals worldwide. It implies that our results may not necessarily be applicable to other institutions. Moreover, adequacy of antimicrobial therapy was defined as a patient's clinical and laboratory improvement after the beginning of the therapy according to the discretion of the providers for each case. Finally, each microbiologic laboratory followed its own protocols and guidelines, not having established a reference laboratory to perform centralized microbiologic analysis of the strains. Therefore, it was not possible to obtain complete and homogeneous susceptibility data.
Conclusion
Prompt identification of patients with cIAIs at high risk of having an infection caused by MDROs is paramount. Our study showed that knowledge of five easily recognizable variables—assessable on hospital admission or as soon as the surgical intervention is concluded—might guide the surgeon to identify patients with cIAIs caused by MDROs, and therefore to choose the most adequate empiric antimicrobial therapy for them. Because only a few international studies have been performed on this topic, we trust that results provided by this study might be valuable for healthcare providers worldwide. These results represent a first step toward further investigations in risk factors for MDROs in patients with a diagnosis of cIAIs.
Footnotes
Author Disclosure Statement
No competing financial interests exist.