Epidemiology and risk factors for nosocomial infection in the respiratory intensive care unit of a teaching hospital in China: A prospective surveillance during 2013 and 2015

Background To determine the epidemiology and risk factors for nosocomial infection (NI) in the Respiratory Intensive Care Unit (RICU) of a teaching hospital in Northwest China. Methods An observational, prospective surveillance was conducted in the RICU from 2013 to 2015. The overall infection rate, distribution of infection sites, device-associated infections and pathogen in the RICU were investigated. Then, the logistic regression analysis was used to test the risk factors for RICU infection. Results In this study, 102 out of 1347 patients experienced NI. Among them, 87 were device-associated infection. The overall prevalence of NI was 7.57% with varied rates from 7.19 to 7.73% over the 3 years. The lower respiratory tract (43.1%), urinary tract (26.5%) and bloodstream (20.6%) infections accounted for the majority of infections. The device-associated infection rates of urinary catheter, central catheter and ventilator were 9.8, 7.4 and 7.4 per 1000 days, respectively.The most frequently isolated pathogens were Staphylococcus aureus (20.9%), Klebsiella pneumoniae (16.4%) and Pseudomonas aeruginosa (10.7%). Multivariate analysis showed that the categories D or E of Average Severity of Illness Score (ASIS), length of stay (10–30, 30–60, ≥60 days), immunosuppressive therapy and ventilator use are the independent risk factors for RICU infection with an adjusted odds ratio (OR) of 1.65 (95% CI: 1.15~2.37), 5.22 (95% CI: 2.63~10.38)), 2.32 (95% CI: 1.19~4.65), 8.93 (95% CI: 3.17~21.23), 31.25 (95% CI: 11.80~63.65)) and 2.70 (95% CI: 1.33~5.35), respectively. Conclusion A relatively low and stable rate of NI was observed in our RICU through year 2013–2015. The ASIS-D、E, stay ≥10 days, immunosuppressive therapy and ventilator use are the independent risk factors for RICU infection.


Background
Nosocomial infection (NI) which also called "hospitalacquired or health care-associated infection" is a serious public health issue affecting hundreds of millions of people every year worldwide [1]. NI is defined as an infection occurring in a patient admitted to the health-care settings for more than 48 but without any evidence that the infection was present or incubating at the time of admission [1][2][3]. In the hospitals or other health care facilities, NI is a leading cause of increased morbidity, mortality and financial burden [1][2][3][4][5][6][7]. The incidence of NI as most studies reporting data ranged from 3.6 to 12% in high-income countries [8][9][10] and 5.7 to 19.1% in low-and middle-income coutries [8,11]. Predisposing factors, i.e., the invasive procedures [12][13][14][15], long hospital stay [16], excessive antibiotics usage [9] and the existence of severe illness [17] lead to NI rate in patients admitted to the intensive care unit (ICU) several fold higher than that in the general hospital population [18][19][20][21]. Now, NI is more concerned as the focus of safety and quality improvements efforts in many hospitals. The study was designed to investigate the epidemiology, risk factors and outcome of NI in a Respiratory ICU (RICU) at the largest teaching hospital in Northwest China.

Study population
This study was conducted in RICU of the First Affiliated Hospital of Xi'an Jiaotong University, which is the largest hospital in Northwest China. It is a 2541-bed teaching hospital with a 16-bed RICU and about 3 millions outpatients annually. The nurse-to-patient ratio in RICU is about 1: 2-3 per shift. A total of 1347 patients admitted to the RICU for more than 48 h were included in the study from January 2013 to December 2015. NI was defined as an infection developed after 48 h of RICU admission and diagnosed according to the the American Center for Disease Control and Prevention (CDC) criteria [22]. In the study, the infection on a different site and with different pathogens from the primary infection that occurred at least 48 h after admission to the RICU was also classified as NI.

Data collection
The patients were followed until discharge from RICU or death, and the information on each patient was recorded on the standard surveillance paper chart. All patients with suspected infection underwent liver and renal function test, whole blood count、urine、fecal and coagulation profile examinations, chest radiography, blood、tracheal aspirate and other body fluids cultures as clinically indicated. Demographic information, i.e., the gender, age, admission and discharge dates, temperature, admission diagnosis, comorbidity, device use and the period of application, laboratory tests, chest radiographs, the isolated pathogens and susceptibility testing to antimicrobial agents, infection sites, drug usage were collected.

The assessment of ASIS
The disease severity was assessed by the Average Severity of Illness Score (ASIS), which was from the Standard for Nosocomial Infection Surveillance of China and established by China Ministry of Health. The criteria of ASIS was as follows: ASIS-A: The patients should be required only routine monitor without intensive care and treatment, and they usually discharged from ICU within 48 h; ASIS-B: The patients, such as the cases admitted to ICU to exclude myocarditis or myocardial infarction, were in stable condition and just required preventive monitor without intensive care and treatment; ASIS-C: The patients, such as those with chronic renal failure, were in stable condition and required intensive care; ASIS-D: The patients in unstable condition but without coma, shock and Disseminated Intravascular Coagulation (DIC), should be performed intensive care and treatment. The treatment should be regularly evaluated and adjusted; ASIS-E: The patients with unstable condition were in coma or shock. The cardio-pulmonary resuscitation, intensive care and treatment should be performed. The intensive care and treatment should be regularly evaluated and adjusted.
According to the Standard for Nosocomial Infection Surveillance of China, the gender, age, admission diagnosis, disease severity, comorbidity, immunosuppressive therapy and invasive procedures were investigated as the potential risk factors for NI in the study.

Research indexes and definitions
The prevalence of nosocomial infection rate was calculated by dividing the total number of nosocomial infections by the total number of patients (× 100). The device-associated nosocomial infection rate was calculated by dividing the total number of device-associated infection by the total days of device application (× 1000). The device utilization (DU) ratios was calculated by dividing the days of device application by the total patient days.

Statistical analysis
Statistical analyses were performed using SPSS 13.0 (serial number 5026743; SPSS Inc., Chicago, IL, USA). Descriptive frequencies were expressed using mean (standard deviation). Chi-square tests were used to compare the rates. For evaluating risk factors of NI, univariate analysis and multivariable logistic regression analysis were used to derive crude OR and adjusted OR, respectively. A p-value < 0.05 was considered statistically significant.

RICU admission patients' characteristics, demographic and clinical data
During the study period, a total of 1347 patients were included, 893 males (66.3%) and 454 females (33.7%), with a mean age of 58.6 years (SD = 17.1). The average length of RICU stay was 8.54 ± 17.72 days, giving 11,501 patient-days. The pneumonia, chronic obstructive pulmonary disease (COPD) and lung cancer accounted for the majority of the RICU admission diagnosis (40.98, 38.9 and 11.6%, respectively). According with ASIS, the patients were mainly in B (42.69%) and C (33.78%) grades. The patients distribution in each month during 2013-2015 was no significant difference (one-way ANVOA, p = 0.064) with 112.2 ± 7.5 numbers per month, the highest and lowest numbers were observed in December (120) and June (100), respectively, Fig. 1a. The COPD exacerbated in December, January and February, pneumonia (community acquired pneumonia) more appeared in July and August, but the proportion of lung cancer in each month was close Fig. 1a. The characteristics of the RICU admission patients were shown in Table 1.

The characteristics of of nosocomial infection in RICU
During the study, 43 of the 552 cases admitted to the RICU with community acquired pneumonia developed NI (a different pathogens than the initial one was isolated). In total, 102 out of 1347 patients experienced NI, 67 males and 35 females, with a prevalence of 7.57% (8.9 per 1000 days). The incidence rate of NI in male (7.5%) was close to that in female (7.7%), p = 0.90. There is no significant change in the incidence rate of NI during the 3 years (range: 7.19 to 7.73%), p = 0.795. The NI in RICU occurred frequently in June, July and August, Fig. 1b.  The NI rate in patients with lung cancer (13.5%) was significantly higher than that in patients with pneumonia (7.9%) and in patients with COPD (6.1%), p = 0.038. With the severity of disease progression from A to E grade, the NI rate increased from 4.07 to 34.78%, p < 0.001, Fig. 2a. The increasing of NI was also found when the length of RICU stay prolonged, p = 0.000, Table 1, Fig. 2b. One hundred seventy-seven pathogens were isolated and identified from the 102 infections, 83 g-negative bacilli and 63 g-positive cocci and 31 fungi. Staphylococcus aureus (20.9%), Klebsiella pneumoniae (16.4%) and Pseudomonas aeruginosa (10.7%) were the most frequently isolated pathogens. The lower respiratory tract, urinary tract and bloodstream accounted for the majority of the RICU-acquired infections (43.1, 26.5 and 20.6%, respectively), Table 2.

Device-associated nosocomial infection in RICU
A total of 87 device-associated nosocomial infections, i.e., 28 catheter-associated urinary tract infections (CAUTI), 12 catheter-associated bloodstream infections (CABSI) and 47 ventilator-associated pneumonia (VAP) were detected in 1347 patients, resulting in an overall rate of 6.5% (7.6 per 1000 days) and accounting for 85.3% of RICU-acquired infections. During the study period, the device application was 3767 days for urinary catheter, 1615 days for central catheter and 4804 days for ventilator, with a device utilization ratio of 0.33, 0.14 and 0.42, respectively. The rate of infection was 9.8 per 1000 days of VAP, 7.4 per 1000 days of CAUTI and 7.4 per 1000 days of CABSI, Table 3. The correlation coefficients between the device utilization and NI were 0.41 for urinary catheter (p = 0.017), 0.139 for central catheter (p = 0.087) and 0.314 for ventilator (p = 0.003). No significant differences were observed between the VAP, CAUTI and CABSI rates (χ 2 = 0.412, P = 0.810).

Risk factors analysis for nosocomial infection in RICU
There are 16 potential risk factors for NI in RICU (Table 4). In the univariate analysis, underlying diseases (lung cancer), ASIS-C˴ D˴ E, RICU stay (≥ 10 days), trauma, diabetes mellitus, immunosuppressive therapy, endotracheal intubation, tracheotomy, utilization of urinary catheter, central catheter and ventilator were identified as risk factors for NI in RICU, P < 0.05. Multivariable logistic regression analysis was conducted to control for the effects of confounding variables. The final analysis showed that ASIS-D˴E, RICU stay (≥ 10 days), immunosuppressive therapy and ventilator utilization are independent risk factors. In RICU ward, patients who were in D˴ E grade, with immunosuppressive therapy, 10-30˴ 30-60 and ≥ 60 days stay and ventilator utilization were 1.65, 5.22, 1.82, 2.32, 8.93, 31.25 and 2.70 times, respectively, more likely to develop NI compared to the control patients who were in A grade, absence of immunosuppressive therapy, with < 10 days stay, and Fig. 2 The comparison of incidence rate of NI for a patients with different ASIS grades and length of RICU-stay, b patients with presence or absence of immunosuppressive therapy, endotracheal intubation, tracheotomy, urinary catheterization, central venous catheterization and ventilator absence of ventilator utilization, respectively, Table 4, Fig. 3. One hundred forty-six patients died during the study period, 21 patients with NI and 125 patients without NI, with a mortality rate of 10.8% (12.7 per 1000 days). The mortality rate in patients with NI was 20.6%, which was significantly higher than that in patients without NI (10.4%), p = 0.001. The incidence of death in patients with NI was 2.32 times to those without NI (95% CI: 1.39-3.89).

Discussion
NI causes increased morbidity, mortality and financial burden at the hospital setting [1][2][3][4][5][6][7]23]. The infection surveillance and risk factors analysis are important prerequisites for the prevention and treatment of NI. At present, abundant literatures focus on the healthcareassociated infection [4,6,[9][10][11]17], infection in ICU [16,[18][19][20][21]23] and device-associated infection [12][13][14][15] have been reported. However, few studies on the topic of infection in RICU have been published. Thus, we conducted this prospective surveillance during 2013 and 2015 to determine the epidemiology and risk factors for NI in RICU at the First Affiliated Hospital of Xi'an Jiaotong University, China. But it was a single cente study and from the largest hospital in Northwest China. The selective bias of the study may affect the generalization of the results.
In our study, there was no significant change in the incidence rate of NI over the 3 years. The overall prevalence of NI in RICU was 7.57%, which was lower than the published rates in European survey (8%) [24] and in India (33.5%) [25]. The mean length of stay was 8.54 days, which was lower than that reported in Italy [26]. In our RICU, COPD was the common underlying diseases, which is in agreement with the published study [26].    [9,16] and Malaysian [27].
In the present study, the device-associated infection accounted for the most of RICU-acquired infections (85.3%). The device utilization ratios (0.14-0.42) were lower than the published rates in Europe, Malaysian and surveys from 61 countries (0.52-0.95) [12,14,[27][28][29]. The VAP rate in our study was significantly lower than that in Greece [14], Malaysian [27]and surveys from 61 countries [28,29] where the rates varied from 13.6 to 20 per 1000 days. The CAUTI rate in our study was lower than that in Malaysian (15.6 per 1000 days) [27], but higher than the published rates (4.2-6.3 per 1000 days)  [14,28,29]. The CABSI rate in the present study was lower than that in Greece (11.8 per 1000 days) [14], but higher than that in Malaysian (3.0 per 1000 days) [27]. Previous studies [11,14,16,17,25,30] indicated that surgery, device utilization, antimicrobial use and length of stay were the risk factors for NI. In our study, the incidence of RICU infection in patients with stay (≥ 10 days), ASIS-C˴ D˴ E, lung cancer, trauma, diabetes mellitus, immunosuppressive therapy, tracheotomy, device utilization was significant higher than that in the control patients (P < 0.05). But only ASIS-D˴ E, stay ≥10 days, immunosuppressive therapy and ventilator utilization are independent risk factors for RICU infection (P < 0.05). The incidence of death in patients with NI was 2.32 times to those without NI.

Conclusions
In conclusion, a relatively low and stable rate of NI was observed in our RICU through year 2013-2015. ASIS-D˴ E, stay ≥10 days, immunosuppressive therapy and ventilator use are independent risk factors for developing infection in our RICU. High mortality rates in patients with infection suggest that infection control activities in RICU must be constantly maintained in order to reduce the rate.