This is a prospective, multi-center, observational study to investigate the epidemiology and outcomes of patients with community-acquired severe sepsis and septic shock who were enrolled from April, 2005 to February, 2009, through the Korean Sepsis Registry System (KSRS) endorsed by the Korean Society of Infectious Diseases. The subjects included adult patients (≥ 18 yr) who had severe sepsis or septic shock within 24 hr of entry to the emergency room or ICU, and excluded those with nosocomial infection as defined elsewhere (11). Patients with health care-associated risk factors who developed severe sepsis or septic shock in the community were also included (12). Only first ICU admission from one patient was included and followed-up until death or hospital discharge.

Eight university hospitals (8 medical and 6 surgical ICUs) located in Seoul were invited and voluntary for participation from the beginning of the study, followed by both medical and surgical ICUs of other 4 university hospitals one year later (Seoul [n = 1], Gyeonggi [n = 2] and Gangwon [n = 1] provinces). The study protocol for enrollment and assessment of the patients were developed by the steering committee of the KSRS that consisted of representative infectious disease specialists.

Infection was defined as microbiologically documented, and classified according to the standard definitions of the Centers for Disease Control and Prevention (11) and/or clinically suspected infection, plus the administration of antibiotics. Primary infection site was assessed by the infectious diseases specialist in each participating hospital, based on the responsible microorganisms isolated and the organ(s) affected. Definitive microorganisms were defined by organisms isolated from a normally sterile site by routine culture, including blood, peritoneal or pleural fluid, cerebrospinal fluid, surgical specimen, or synovial fluid, or a bacterial pathogen isolated from cultures of respiratory specimens or a urine specimen with bacterial growth of ≥ 1 × 10³ colony-forming units/mL. Sepsis and sepsis-related conditions were diagnosed in accordance with the criteria proposed by ACCP/SCCM (American College of Chest Physician/Society of Critical Care Medicine) (1). Definition of severe sepsis was the presence of sepsis plus at least one organ or system dysfunction. The criteria for acute organ or system dysfunction were modified from previous criteria (13): cardiovascular dysfunction (systolic blood pressure ≤ 90 mmHg or mean arterial pressure ≤ 70 mmHg that responds to administration of intravenous fluid), renal dysfunction (urine output < 0.5 mL/kg/hr for 1 hr despite adequate fluid resuscitation or serum creatinine ≥ 177 µM not attributable to chronic kidney failure), respiratory dysfunction (ratio of partial arterial oxygen tension to inspired fractional oxygen ≤ 200), hematologic dysfunction (platelet < 100 × 10³/µL or platelet decreased to < 50% in the 2 days preceding admission or INR (International Normalized Ratio) > 1.5 in the absence of systemic anticoagulant agents), hepatic dysfunction (bilirubin > 43 µM or alanine transaminase > 50 IU/L), central nervous system dysfunction (documentation of deterioration in mental status within 24 hr before admission), gastrointestinal dysfunction (active gastrointestinal tract bleeding requiring transfusions within 24 hr of onset of sepsis), metabolic dysfunction (arterial pH ≤ 7.30 or the base deficit ≥ 5.0 mM in association with a plasma lactate > 2 times the upper limit of the normal value). Septic shock was defined as either systolic blood pressure ≤ 90 mmHg or mean arterial pressure ≤ 70 mmHg for at least 1 hr after adequate fluid resuscitation or use of vasopressors to maintain systolic blood pressure > 90 mmHg or mean arterial pressure > 70 mmHg.

Data from each patient were collected on standardized web-based case report form by the physicians or research nurses in each hospital. They were provided detailed instructions explaining the aim of the study, instructions for data collection, and definitions for variables by operating manual before starting data collection and throughout the study period. The coordinating center at the Korea University Anam Hospital supported data registration at each center, and all data were checked to be within acceptable ranges.

Data collection on admission included demographic data, presence of underlying comorbidity, Charlson's index score (14). Acute Physiology and Chronic Health Evaluation (APACHE) II score (15), Sequential Organ Failure Assessment (SOFA) score (16), microbiology and clinical infections were collected. We recorded the admission date to ICU and the date of in hospital-death or hospital discharge. Sepsis-related mortality was judged by investigators in the participating hospitals.

Descriptive analysis was performed. Differences of categorical variables were compared using the chi-square test or Fisher exact test. Continuous variables were expressed as mean ± standard deviation (SD), and Student's two-tailed t test and two-way ANOVA were used for comparing means. To determine the effect of gender on 28-day and in-hospital mortality, multiple logistic regression analysis with forward variable selection method was done. The all variables significant in univariate analysis (P < 0.05) were included in the multivariate model. Odds ratio (OR) and 95% confidence interval (CI) were calculated according to standard methods after adjusting for confounding factors. All tests were two-tailed and a P value < 0.05 was considered significant. Data were analyzed using SPSS version 13.0 for Windows (SPSS, Chicago, IL, USA).

The study protocol was approved by the institutional review board of the Korea University Anam Hospital (IRB No. ED 0513). Since this observational study required no deviation from routine medical practice, informed consent was waived by board.

During the study period, a total of 1,192 patients met the enrollment criteria for the community-acquired severe sepsis or septic shock. The demographic and baseline clinical characteristics of patients by gender are shown in Table 1. Men accounted for 55% (656/1,192), and had lower mean age than in women (63.9 ± 14.0 yr vs 66.4 ± 14.8 yr; P = 0.003). The peak of age distribution was 60-69-yr-of-age.

Any type of comorbidity was present in 847 (71.1%) patients. Men had higher rates of central nervous system diseases, cancer, liver diseases and lung diseases. However, cardiovascular diseases and connective tissue diseases were higher in women. The severity of comorbidity as measured by Charlson's index score was also greater in men.

The most common primary site of infection was the respiratory system, followed by urinary tract, abdomen, skin and soft tissue and primary bloodstream. Men were more likely to have respiratory (39.2% vs 19.2%, P < 0.001), abdominal (26.8% vs 20.9%, P = 0.017) and skin and soft tissue infections (11.0% vs 4.7%, P < 0.001), compared to women. Women showed higher rates of urinary tract infections (11.4% vs 44.8%, P < 0.001).

Clinical infections with identification of pathogens were documented in 831 (69.7%) patients. Distribution of microorganisms and sites of isolation were shown in Table 2. Gram negative bacteria were the predominant organisms of the isolates, accounting for 43.0% (n = 513), while gram positive bacteria accounted for 20.4% (n = 243). Polymicrobial infections accounted for 3.6%, followed by fungi (0.7%) and anaerobes (0.6%). Escherichia coli (22.1%) was the most common isolate, followed by Klebsiella species (12.6%) and Staphylococcus aureus (8.4%). Bacteremia developed in 35.4% (422/1,192) of patients, with the frequency of E. coli (37.7%; 159/422), Klebsiella species (22.0%; 93/422), and S. aureus (11.8%; 50/422). Based on clinical samples for the bacterial isolates, Streptococcus pneumoniae, S. aureus and Klebsiella species were more frequently isolated from respiratory samples. E. coli and Klebsiella species were the most frequent isolates from abdominal aspirates. Gram positive bacteria were more common in male patients than in female patients (25.3% vs 14.4%, P<0.001), whereas gram negative organisms were more common in female patients (36.6% vs 50.9%, P<0.001).

Clinical severity and outcomes of the 1,192 patients analyzed are shown in Table 3. The mean APACHE II score of patients on ICU admission was 18.8±7.3. The mean number of organ dysfunctions within 24 hr after admission was 2.6±1.6. The hematologic dysfunction was the most frequent, acute organ dysfunction, followed by hepatic, renal, respiratory, metabolic, cardiovascular, and gastrointestinal dysfunction. The mean SOFA score was 7.5 points on day 1, and decreased to 5.0 points on day 5. There were significant differences between men and women in the mean APACHE II score, the mean SOFA score, and in occurrence of acute organ dysfunction. Men were more likely to have renal, respiratory, hepatic, and metabolic organ dysfunction, compared to women (P<0.05).

Septic shock was present in 62.1% (740/1,192) of patients. There was no difference in the presence of septic shock between men and women. Development of septic shock was significantly associated with source of primary infection; abdomen (OR, 1.31; 95% CI, 0.96-1.77, P=0.088), bloodstream (2.24; 1.12-4.49, P=0.023), respiratory system (0.64; 0.48-0.84, P=0.001).

The overall 28-day, in-hospital, and sepsis-related mortality rates were 23.0% (n=274), 28.0% (n=334), and 23.6% (n=281), respectively (Table 3). The patients with septic shock had significantly higher mortality rates compared to those without septic shock: the 28-day mortality rate (26.1% vs 17.9%, P=0.001), in-hospital mortality rate (30.8% vs 23.5%, P=0.006), and sepsis-related mortality rate (27.2% vs 17.7%, P<0.001). There are also significant differences between men and women in 28-day, in-hospital, and sepsis-related mortality rates. Men had significantly higher mortality in all of the age groups compared to women (Fig. 1).

The mean length of hospital stay for survivors and non-survivors was 26.5±34.5 days and 21.1±45.7 days, respectively. However, there was no significant gender difference in length of hospital stay for survivors or non-survivors.

Multiple logistic regression analysis was performed to identify the risk factors which could explain the differences of the 28-day mortality and in-hospital mortality between men and women. Independent variables that were significantly different between men and women in univariate analysis were included in the models. The significant variables selected in the final model of the 28-day mortality were urinary tract infection (OR, 0.33; 95% CI, 0.17-0.59), APACHE II score (1.06; 1.02-1.09), SOFA score on day 1 (1.11; 1.04-1.18) and metabolic dysfunction (2.09; 1.36-3.22). In the model of in-hospital mortality, cancer (1.89; 1.13-3.17), urinary tract infection (0.25; 0.13-0.46), APACHE II score (1.05; 1.02-1.09), SOFA score on day 1 (1.13; 1.06-1.21) and metabolic dysfunction (2.24, 1.45-3.45) were significant variables selected (Table 4). However, both age and gender were not significant variables.