Study design

The aim of the CIAO Study was to describe the epidemiological, clinical, microbiological, and surgical treatment profiles of community-acquired and healthcare-associated complicated intra-abdominal infections (IAIs) based on data collected over a 6-month period (January-June 2012) from 68 medical institutions throughout Europe (see Figure 1).

Patients with either community-acquired or healthcare-associated complicated intra-abdominal infections (IAIs) were included in the study.

The center coordinator of each participating medical institution collected and compiled clinical data in an online case report database.

The collected data included the following: (i) patient and disease characteristics, i.e. patient demographic data, type of infection (nosocomial or community-acquired), severity criteria, and previous antibiotic therapy administered in the 7 days preceding surgery; (ii) origin of infection, surgical procedures performed, and antibiotic therapies administered; and (iii) microbiological data, i.e. identification of bacteria and microorganismal pathogens within the peritoneal fluid, the identification of yeasts (if present), and the antibiotic susceptibilities of bacterial isolates.

This observational study did not attempt to change or modify the laboratory or clinical practices of the participating physicians or their respective institutions, and it did not require informed consent or formal approval by an Ethics Committee.

A Scientific Committee was established to impartially assess the objectives, methodology, and overall scientific quality of the project.

The study was monitored by the coordination center, which processed and verified missing or unclear data submitted to the central database.

Statistical analysis was performed using STATA® statistical software.

Patients

2,152 patients with a mean age of 53.8 years (range 4–98) were enrolled in the CIAO Study. 996 patients (46.3%) were women and 1,156 (53.7%) were men. Among these patients, 1,701 (79%) were affected by community-acquired IAIs while the remaining 451 (21%) suffered from heathcare-associated infections. Intraperitoneal specimens were collected from 1,338 (62.2%) of the enrolled patients.

787 patients (36.5%) were affected by generalized peritonitis while 1,365 (63.5%) suffered from localized peritonitis or abscesses.

282 patients (13.1%) were admitted in critical condition (severe sepsis/septic shock).

Source control

The open appendectomy was the most common means of addressing complicated appendicitis. 443 patients (55.1%) admitted for complicated appendicitis underwent open appendectomies: 343 patients (77.4%) for localized infection or abscesses and 100 patients (29.1%) for generalized peritonitis. A laparoscopic appendectomy was performed for 318 patients (39.8%) with complicated acute appendicitis; of these patients, 217 underwent the procedure for localized peritonitis/abscesses and 101 underwent the procedure for generalized peritonitis. Open bowel resection was performed for 7 patients affected by complicated appendicitis. In the other 30 cases of complicated appendicitis (4.3%), conservative treatment (percutaneous drainage, surgical drainage, and non-operative treatment) was performed. 1.6% of patients underwent percutaneous drainage and interval appendectomies to address appendicular abscesses.

Among the patients with complicated cholecystitis (289), the open cholecystectomy was the most frequently performed procedure. 48.4% and 40.8% of cholecystitis patients underwent open and laparoscopic cholecystectomies, respectively. The remaining patients were treated with conservative methods (percutaneous drainage, non-operative treatment).

Among the patients with complicated diverticulitis (166) the Hartmann resection was the most frequently performed procedure. 73 patients (43.2%) underwent a Hartmann resection, and of these resections, the vast majority were open procedures (94.5% open compared to 5.5% laparoscopic). 54 of these patients (74%) underwent a Hartmann resection for generalized peritonitis, while the remaining 19 (26%) underwent the same procedure for localized peritonitis or abscesses. Colo-rectal resection was performed in 41 cases (24.7%). Laparoscopic resection was performed for only 3 patients (2 patients with and 1 patient without protective stoma) while open resection was performed for 38 patients (27 with and 11 without protective stoma).

The remaining patients received conservative treatment (percutaneous drainage, non-operative treatment, surgical drainage and stoma). 11 patients underwent laparoscopic drainage.

For patients with gastro-duodenal perforations (156 cases), the most common surgical procedure was gastro-duodenal suture. 107 patients underwent open gastro-duodenal suture (68.6%) and 18 patients underwent laparoscopic gastro-duodenal suture (11.5%). 16 patients (10.3%) underwent gastro-duodenal resection and 16 patients (10.3%) received conservative treatment (non-operative treatment, surgical drainage). The remaining patients underwent alternative procedures.

Of the 100 patients with small bowel perforations, 83 underwent open small bowel resection (83%) and 3 (3%) underwent laparoscopic small bowel resection. The remaining 14 patients (14%) were treated non-surgically.

Among the 158 patients with colonic non-diverticular perforation, 52 (32.9%) underwent open Hartmann resection, 55 (34.8%) underwent open resection with anastomosis and without stoma protection, and 23 underwent open resection with stoma protection (14.6%).

369 cases (17.1%) were attributable to post-operative infections. Anastomotic leaks were the most prevalent cause of post-operative infection. Of all post-operative infections, 40.2% resulted from colo-rectal leaks, 32.1% from upper gastro-intestinal leaks, 14.5% from biliary leaks, 11.2% from pancreatic leaks, and 1.9% from urinary leaks.

Source control was successfully implemented for 1,985 patients (92%) and proved ineffective for 167 patients (8%).

Microbiology

Intraperitoneal specimens were collected from 1,339 patients (62.2%).

These specimens were obtained from 977 of the 1,701 patients presenting with community-acquired intra-abdominal infections (57.4%).

Intraperitoneal specimens were collected from 362 (80.3%) of the remaining 451 patients with nosocomial intra-abdominal infections.

The major pathogens involved in intra-abdominal infections were found to be Enterobacteriaceae.

In community-acquired IAIs, Extended-Spectrum Beta-Lactamase (ESBL)-producing Escherichia coli isolates comprised 10.1% (64/632) of all Escherichia coli isolates, while ESBL-positive Klebsiella pneumoniae isolates represented 33.9% (37/109) of all Klebsiella pneumoniae isolates.

ESBL-positive Enterobacteriaceae were more prevalent in patients with nosocomial IAIs than they were in patients with community-acquired IAIs. ESBL-positive Escherichia coli isolates comprised 22.4% (34/152) of all identified Escherichia coli isolates, while ESBL-positive Klebsiella pneumoniae isolates made up 50% (26/52) of all identified Klebsiella pneumoniae isolates.

There were 5 isolates of Klebsiella pneumoniae resistant to Carbapenems. All Carbapenem-resistant Klebsiella pneumoniae isolates were acquired in an intensive care setting.

Among the identified aerobic gram-negative isolates, there were 80 isolates of Pseudomonas aeruginosa, comprising 5.3% of all identified aerobic bacteria isolates (4.3% in patients with community-acquired infections versus 6.7% in patients with nosocomial infections).

The 3 Pseudomonas aeruginosa strains resistant to Carbapenems were also obtained from nosocomial infections.

Among the identified aerobic gram-positive bacteria, Enterococci (E. faecalis and E. faecium) were the most prevalent, representing 16% of all aerobic isolates, and were identified in 241 cases. 22 glycopeptide-resistant Enterococci were identified; 16 were glycopeptide-resistant Enterococcus faecalis isolates and 6 were glycopeptide-resistant Enterococcus faecium isolates.

Although Enterococci were also present in community-acquired infections, they were far more prevalent in nosocomial infections.

Tests for anaerobes were conducted for 680 patients.

197 anaerobes were observed. The most frequently identified anaerobic pathogen was Bacteroides. 126 Bacteroides isolates were observed during the course of the study. Among the Bacteroides isolates, there were 3 Metronidazole-resistant strains.

Outcome

The overall mortality rate was 7.6% (163/2,152).

521 patients (24.2%) were admitted to the intensive care unit in the early recovery phase immediately following surgery.

255 post-operative patients (11.8%) ultimately required additional surgeries; 66.7% of follow-up laparotomies were unplanned “on-demand” procedures and 20% were anticipated surgeries. Overall, 11.3% of these patients underwent open abdominal procedures.

For healthcare associated infections, the setting of acquisition was also a variable found to be predictive of patient mortality (chronic care setting: OR=5.2; 95%CI=1.7-8.4; p<0.0001, non-operative hospital setting: OR=3.8; 95%CI=2.4-10.9; p<0.0001, and post-operative hospital setting: OR=2.5; 95%CI=1.7-3.7; p<0.0001).

Among the various sources of infection, colonic non-diverticular perforation (OR=117.4; 95%CI=27.9-493.9, p<0.0001), complicated diverticulitis (OR=45.4; 95%CI=10.4-198.6; p<0.0001), and small bowel perforation (OR=125.7; 95%CI=29.1-542; p<0.0001) were significantly correlated with patient mortality.

Mortality rates did not vary to a statistically significant degree between patients who received adequate source control and those who did not. However, a delayed initial intervention (a delay exceeding 24 hours) was associated with an increased mortality rate (OR=2.6; 95%CI=1.8-3.5; p<0.0001).

The nature of the immediate post-operative clinical period was a significant predictor of mortality (severe sepsis: OR=33.8; 95%CI=19.5-58.4; p<0.0001, septic shock: OR=59.2; 95%CI=34.4-102.1; p<0.0001). Patients requiring ICU admission (OR=18.6; 95%CI=12-28.7; p<0.0001) were also associated with increased mortality rates.

WBC counts greater than 12,000 or less than 4,000 (OR=2.8; 95%CI=1.8-4.4; p<0.0001), and core body temperatures greater than 38°C or less than 36°C (OR=3.3; 95%CI=2.2-5; p<0.0001) by the third post-operative day were significant predictors of patient mortality.

Source control

Complicated intra-abdominal infections are an important source of patient morbidity and are frequently associated with poor clinical prognoses, particularly for patients in high-risk categories.

The CIAO Study has confirmed that acute appendicitis is the most common intra-abdominal condition requiring emergency surgery in Europe. Both open and laparoscopic appendectomies are viable treatment options for complicated appendicitis [4]. The laparoscopic appendectomy is a safe and effective means of surgical treatment for addressing complicated intra-abdominal infections, but open surgery still retains several clinical advantages, including a reduced probability of post-operative intra-abdominal abscesses [5]. CIAO Study data indicate that the open approach was used in 55.1% of complicated appendicitis cases while the laparoscopic approach was performed in 39.8% of these cases.

For patients with periappendiceal abscesses, the proper course of surgical treatment remains a point of contention in the medical community. However, this contention notwithstanding, the most commonly employed treatment appears to be drainage with subsequent appendectomy [6].

Although guidelines for the management of intra-abdominal infections commonly assert that patients with periappendiceal abscesses should be treated with percutaneous image-guided drainage, few patients underwent this procedure.

The laparoscopic versus open cholecystectomy debate has been extensively investigated in recent years. In the CIAO Study, the open cholecystectomy was the most common means of treating cholecystitis; 48.4% of patients with complicated cholecystitis underwent this procedure. By contrast, 118 patients (40.8%) underwent the laparoscopic procedure.

The optimal surgical management of colonic diverticular disease complicated by peritonitis remains a controversial issue in the medical community.

Hartmann’s resection has historically been considered the procedure of choice for patients with generalized peritonitis and continues to be a safe and reliable technique for performing an emergency colectomy in the event of perforated diverticulitis, particularly in elderly patients with multiple co-morbidities [7–10].

More recently, however, reports have suggested that primary resection and anastomosis may be the optimum approach to addressing diverticulitis, even in the presence of diffuse peritonitis [11].

According to CIAO Study data, the Hartmann resection was the most frequently performed procedure to address complicated diverticulitis in Europe. 43.2% of patients underwent a Hartmann resection, and of these resections, the vast majority were open procedures (94.5% open compared to 5.5% laparoscopic). 54 of these patients (74%) underwent a Hartmann resection for generalized peritonitis, while the remaining 19 (26%) underwent the same procedure for localized peritonitis or abscesses.

22.5% of patients underwent colo-rectal resection to address complicated diverticulitis.

Microbiology

The significance of microbiological analysis of infected peritoneal fluid in community-acquired intra-abdominal infections has been debated in recent years.

Cultures from the site of infection should always be obtained for patients with nosocomial infections as well as for patients with community-acquired infections who are known to be at risk for drug-resistant strains. In these patients, causative pathogens and resistance patterns are unpredictable and always require cultures from the site of infection [4].

Bacterial cultures and analyses may be often clinically superfluous, particularly when the etiological agents are readily predictable [12]. However, some authors maintain that in-depth bacterial diagnosis has practical significance, even in low-risk patients with community-acquired IAIs. They argue that this analysis plays an important role in documenting epidemiological shifts in antimicrobial resistance patterns associated with community-acquired IAIs and in guiding individualized follow-up therapy. For high-risk patients with community-acquired IAIs or in the event of nosocomial IAIs, clinicians should always obtain cultures from the site of infection.

According to CIAO Study data, intraperitoneal specimens were collected from 62.2% of patients; these samples were obtained from 57.4% of patients with community-acquired IAIs and from 80.3% of patients with nosocomial IAIs.

In many clinical laboratories, species identification and susceptibility testing of anaerobic isolates are not routinely performed [13].

Of the total patients tested for aerobic microorganisms, 42.9% underwent tests for anaerobes.

The major pathogens involved in community-acquired intra-abdominal infections are Enterobacteriaceae, Streptococcus sp ecies, and certain anaerobes (particularly B. fragilis). Compared to community-acquired infections, nosocomial infections typically involved a broader spectrum of microorganisms, encompassing ESBL-producing Enterobacteriaceae, Enterococcus, Pseudomonas, and Candida species in addition to the Enterobacteriaceae, Streptococcus species, and anaerobes observed in community-acquired IAIs.

Antimicrobial resistance has become a major challenge complicating the treatment and management of intra-abdominal infections.

The main resistance threat is posed by ESBL-producing Enterobacteriaceae, which are becoming increasingly common in community-acquired infections.

Many factors can increase the prevalence of ESBL activity in community-acquired intra-abdominal infections, including excessive use of antibiotics, residence in a long-term care facility, and recent hospitalization. Further, male patients and patients over the age of 65 appear to be particularly susceptible to ESBL-producing bacterial infections [14].

According to CIAO Study data, ESBL producers were the most commonly identified drug-resistant microorganism involved in IAIs.

Recent years have seen an escalating trend of Klebsiella pneumoniae Carbapenemase (KPC) production, which continues to cause serious multidrug-resistant infections around the world. The recent emergence of Carbapenem-resistant Enterobacteriaceae is a major threat to hospitalized patients.

In addition to hydrolyzing Carbapenems, KPC-producing strains are also resistant to a variety of other antibiotics, and consequently, these infections pose a considerable challenge for clinicians in acute care situations.

KPC-producing bacteria are most common in nosocomial infections, particularly in patients with previous exposure to antibiotics [15].

5 identified isolates of Klebsiella pneumoniae proved resistant to Carbapenems, and each was acquired in an intensive care setting.

The rate of Pseudomonas aeruginosa among aerobic isolates was 5.2%. There was no statistically significant difference in Pseudomonas prevalence between community-acquired and nosocomial IAIs.

Enterococci (E. faecalis and E. faecium) were identified in 15.7% of all aerobic isolates.

Although Enterococci were also identified in community-acquired infections, they were far more prevalent in nosocomial infections.

In the CIAO Study, 138 Candida isolates were observed among 1,890 total isolates (7.3%). Candida prevalence was significantly higher in the nosocomial IAI group than it was in the community-acquired IAI group.

Outcome

Of the 2,152 patients enrolled in the study, there were 163 deaths (7.6%).

According to univariate statistical analysis of the data, critical clinical condition of the patient upon hospital admission (defined by severe sepsis/septic shock) as well as critical clinical condition in the immediate post-operative period and ICU admission were all significant risk factors predictive of patient mortality. WBCs greater than 12,000 or less than 4,000 and core body temperatures greater than 38°C or less than 36°C by the third post-operative day were predictors of patient mortality. Among the various sources of infection, colonic non-diverticular perforations, complicated diverticulitis, and small bowel perforations correlated strongly with patient mortality.

Mortality rates did not vary to a statistically significant degree between patients who received adequate source control and those who did not. However, a delayed initial intervention (a delay exceeding 24 hours) was associated with an increased mortality rate.

According to stepwise multivariate analysis (PR=0.005 and PE=0.001), several criteria were found to be independent variables predictive of patient mortality, including patient age, the presence of an intestinal non-appendicular source of infection (colonic non-diverticular perforation, complicated diverticulitis, small bowel perforation), a delayed initial intervention (a delay exceeding 24 hours), sepsis and septic shock in the immediate post-operative period, and ICU admission.