Multi-center prospective study on central line-associated bloodstream infections in 79 ICUs of China

Background China has not yet established a national surveillance network such as NHSN from America, so there is still no large-scale investigations on central line-associated bloodstream infection (CLABSI) incidence. Several retrospective studies in China reported that the incidence of CLABSI varied due to inconsistent diagnostic criteria. We performed a nationwide survey to investigate the utilization rate of central venous catheters (CVCs) and the incidence of CLABSI in ICUs of different areas of China. Methods This is a prospective multi-center study. Patients admitted to ICUs with the use of CVCs between January 1, 2014 and December 31, 2018 were enrolled in this study. Hospitals were given the definition of catheter-related bloodstream infection as: a laboratory-confirmed bloodstream infection where CVC was in place on the date of event or the day before. The characteristics of patients, information of catheterization, implementation rates of precautions, and CLABSIs were collected. The statistical analysis was performed by SPSS 25.0 software and website of Open Source Epidemiologic Statistics for Public Health. Results A total of 38,212 patients and 466,585 catheter days were involved in surveillance. The average CLABSI incidence in a thousand catheter days was 1.50, the lowest incidence unit was in pediatric ICU (0/1000 catheter days), and the lowest incidence area was in Northeast China (0.77/1000 catheter days), while the highest incidence unit was in cardiac ICU (2.48/1000 catheter days) and the highest incidence area was in Eastern China (1.62/1000 catheter days). The average utilization rate of CVC was 42.85%, the lowest utilization rate was in pediatric ICU (5.85%) and in Central China (38.05%), while the highest utilization rate was in surgical ICU (64.92%) and in Western China (51.57%). Among the 702 CLABSI cases reported, a total of 735 strains of pathogens were cultured. Staphylococcus spp. was the most common organism isolated (27.07%), followed by Enterobacteriaceae (22.31%). The implementation rates of all precautions showed an upward trend during the study period (P ≤ 0.001). Conclusion The average incidence of CLABSI in ICUs in China is 1.5/1000 catheter days, similar to the rates reported in developed countries but lower than previous reports in China. CLABSI incidence showed regional differences in China. It is necessary to implement targeted surveillance of CLABSI cases by using standardized CLABSI surveillance definitions and methodologies.


Background
With the development of medical science and technology, various vascular catheters have been widely used in clinical practice, not only in critical wards and hemodialysis departments, but also in general departments, mainly for infusion, chemotherapy, hemodialysis, parenteral nutrition support and hemodynamic surveillance [1]. However, the wide use of catheters increases the risk of catheter-related local and bloodstream infections in patients. Mortality induced by central line-associated bloodstream infection (CLABSI) could be as high as 28-30% [2].
Several retrospective studies in China reported that the incidence of CLABSI varied from 1.67 to 16.96 per 1000 catheter days [3][4][5], and most common pathogens associated with CLABSI were Gram-positive bacteria [6,7]. The wide range of the reported incidences mainly caused by the inconsistent diagnostic criteria and different capacity of diagnosis. To our knowledge, there is still no large-scale investigations on CLABSI incidence in mainland China, due to the difficulties in data collection and low compliance of surveillance. Therefore, we performed a nationwide survey of CLABSI incidence in intensive care unit (ICU) to evaluate the current status in China and develop strategies to reduce CLBASI rates.

Study design
This study conducted a prospective surveillance of CLABSI infection in ICU through multi-center cooperation.

Site of study
Data collection was based on province/municipality as a basic unit, and the participating hospitals and ICUs had changed slightly every year during the study.
Totally 79 ICUs (50 General ICUs, 6 Emergency ICUs, 6 Surgical ICUs, 5 Neurological ICUs, 4 Respiratory ICUs, 4 Medical ICUs, 2 Coronary Care Units, 2 Pediatric ICUs) from 62 hospitals in ten provinces (Liaoning, Shandong, Henan, Shanxi, Jiangsu, Zhejiang, Guangdong, Guizhou, Sichuan, Hunan) and one municipality (Beijing) participated in this study. Only one of the 62 hospitals in this survey is a special hospital (Maternal and Child Health Hospital), and the rest are general hospitals, all participating hospitals have more than 500 beds.

Study period
The study was from January 1, 2014 to December 31, 2018.

Inclusion criteria
Patients with CVCs admitted to the participating ICUs or patients insert new CVCs in the participating ICUs between January 1, 2014 and December 31, 2018. Only the subclavian vein, internal jugular vein, and femoral vein were considered great vessels for the purpose of reporting CLABSI in this study.

Exclusion criteria
Patients stayed in ICU less than 48 h.

Follow-up periods
Patients were followed till 48 h after discharge/transfer out from the ICU, or the death.

Active surveillance
Daily surveillance were taken by designated nurses from ICU to record the number of hospitalized patients and the number of patients with CVC every day at 8:00 am. The staffs of infection control department recorded the patients' demographic details, reason for admission, outcome, information of catheterization and CLABSI, made spot checks and recorded the implementation of precautions.
Hospitals were given the definition of catheter-related bloodstream infection as: a laboratory-confirmed bloodstream infection where CVC was in place on the date of event or the day before. Laboratory-confirmed bloodstream infection criteria were consistent to the United States CDC's NHSN 2013 surveillance criteria [8].
Since April 1, 2014, the participating hospitals implemented observation, education, performance monitoring, feedback to continuously promote the adherence rates of following intervention:hand hygiene, using fullbarrier precautions during the insertion of central venous catheters, cleaning the skin with chlorhexidine, avoiding the femoral site if possible, and removing unnecessary catheters.
Because this is a multi-center cooperative study, it is especially important to ensure the consistency of the surveillance. The surveillance staffs of each hospital were trained before the investigation, about the surveillance procedure, diagnostic criteria and questionnaire filling requirements. Moreover, quality supervisor of each province/municipality would collect and review provincial data every month during the study period to ensure the completeness and credibility of the study. All data were submitted via a uniform data collection tool, with electronic data submission available from the start of the study.

Statistical analysis
Statistical analysis was performed using SPSS version 25 and website of Open Source Epidemiologic Statistics for Public Health. Medians and interquartile ranges were used for continuous variables. Linear-by-Linear Association was used to compare the trend of annual CLABSI incidences and annual implementation rates of precautions. The two incidences of CLABSI from different regions and ICUs were compared by using Person-time Analysis.

Results
Total 38,212 patients were admitted to the ICUs with 45,978 CVCs utilized in the study period. The surveillance data were reported over 466,585 catheter days. 702 CLABSI cases occurred, the average CLABSI incidence in a thousand catheter days was 1.50 and the average CVC utilization rate was 42.85%. As shown in Table 1, the participating hospitals and ICUs had changed slightly every year during the study, and total 79 ICUs from 62 hospitals in 11 provinces and municipalities participated in this study. Table 2 showed the characteristics of patients and catheter enrolled in this study. The age of patients ranged between 1 days to 104 years old. Only 37,770 catheters insert in the participated hospital (25,113 catheters from ICUs, and 12,657 from non-ICU departments, such as Emergency Departments, Operating Rooms, etc.). Multi-position catheterization were used in 690 patients (1.81%) which were significantly more frequently seen with CLABSI.

Rate of CVC utilization and CLABSI incidence
During the study period, the utilization rate of CVC had slowly decreased to about 40% yearly, but the CLABSI incidence in a thousand catheter days had increased to about 1.65 ( Table 3). The implementation rates of all precautions showed an upward trend during the study period (P ≤ 0.001) ( Table 4).
Rate of CVC utilization and CLABSI incidence in different kinds of ICU were shown in Table 5. The department with the lowest CVC utilization rate was pediatric ICU, the department with the highest CVC utilization rate was surgical ICU, and there was no infection case monitored in pediatric ICU. The lowest and highest CLABSI incidence in different kind of ICU was pediatric ICU (0/1000 catheter days) and cardiac ICU (2.48/1000 catheter days), respectively.
Rate of CVC utilization and CLABSI incidence in different areas of China were shown in Table 6. The area with the lowest CLABSI incidence and CVC utilization rate was Northeast China (0.77/1000 catheter days) and Central China (38.05%), respectively, and the highest CLABSI incidence and CVC utilization rate during the study period was Eastern China (1.62/1000 catheter days) and Western China (51.57%), respectively.

Microbiology and susceptibility patterns
Among 702 CLABSI cases reported in the ICUs, 735 pathogens were isolated.     [10]. A surveillance of CLABSI incidence in ICUs of four European countries showed that CLABSI incidence was 4.2/1000 catheter days in UK, 1.5/1000 catheter days in Germany, 2.0/1000 catheter days in Italy, and 1.23/1000 catheter days in France [11]. 384 CLABSI events were reported over 303,968 CVC days, corresponding to a rate of 1.26/1000 catheter days (95% confidence interval, 1.14-1.40) in Australian [12]. Studies in 18 developing countries showed that CLABSI incidence was around 6.9-15.2 per 1000 catheter days [13]. In this study, CLABSI incidence was comparable to some developed countries, but was significantly lower than developing countries including China [14]. Since most of our data were collected from tertiary care and general hospitals with more than 500 beds (only three secondary care hospitals) in 11 provinces and municipality, our results reflected the current incidence of CLABSI and adherence rates of interventions in large general hospitals in China.
The average utilization rate of CVC in this study was lower than that reported in 140 hospitals of Jiangsu province in China (52.27%) [15] and that reported in Germany (82.2%) [16]. The difference may be related to the fact that we only included central line inserted in internal jugular vein, subclavian vein, and femoral vein, but excluded central line inserted in other venous and arterial such as implantable venous access port or PICC. It can also be seen from our results that the utilization rate of CVC of different ICUs were quite different, and our study showed a low percentage of surgical ICU which had higher utilization rate of CVC.
Although we have developed detailed protocols to reduce the incidence of CLABSI, CLABSI incidence Other Gram-negative bacilli 9 (1.22) had no significant decrease during the study period, but the utilization rate of CVC had slowly decreased to about 40%. From the 3rd year of the study, the research group adjusted the types of participating ICU with focus on general ICU, medical ICU and respiratory ICU, which had relatively high utilization rate and high CLABSI incidence. This may explain the fluctuations of CLABSI incidence annually. Pediatric ICU had the lowest utilization rate because pediatrics choose PICC as the main way of infusion, and there was no infection case for short surveillance time, so the infection rate was the lowest. In contrast, most patients in surgical ICU need CVCs due to surgery, leading to the highest utilization rate. Among different ICUs, general ICU had the maximum sample size and the most stable surveillance source, so the data were more reliable. CLABSI incidence of general ICU was the closest to the average incidence because the patients came from various specialist departments and had more complicated condition. Other ICUs such as emergency ICU, medical ICU, respiratory ICU and neurological ICU had relatively lower incidences of CLABSI.
In addition, CVC utilization rate and CLABSI incidence showed differences across different regions of China. Infections tend to occur in the eastern part of the country, with relatively fewer infections in the northeast part. Due to current significant imbalance in medical and health resources, some patients, especially the critically ill, will choose to go to a large hospital in the eastern region for medical treatment. This may increase the number of CLABSI cases in the eastern region to some extent. The CVC utilization rate in Western part of China was higher than in other regions, and the rate in some hospitals and ICUs was even as high as 90%, indicating that some hospitals have placed CVCs for patients who do not meet the indications. Nevertheless, the differences between regions cannot be explained only by the aforementioned parameters, and need more detailed studies.
We also investigated the pathogenic microorganisms and their drug resistance associated with CLABSI cases. Studies have shown that the main pathogens of CLABSI are coagulase-negative Staphylococcus, S aureus, enterococcus, and candida due to the adhesion to the epidermis and wounds [6,12,[17][18][19][20]. However, recent studies have found that the incidence of CLABSI caused by Gramnegative bacilli is increasing [7,[21][22][23][24][25]. The results of this study showed that Gram-negative bacteria account for a greater proportion (49.24%) than Gram-positive bacteria (38.17%) in the CLABSI cases, and there is an upward trend. Participating hospitals have adopted CLABSI preventive measures, such as cluster care strategies, which have reduced the adhesion of Gram-positive bacteria to the epidermis and wounds. This may explain the decline in the proportion of Gram-positive bacteria in CLABSI pathogens.
As a morbidity survey, our study has several limitations. First, the areas covered by this study is not complete. There are 34 provinces, municipalities and autonomous regions in China, and our study only covered 11 of them. In particular, the coverage of the western and northeastern regions is not enough, such as Xinjiang and Tibet, which are precisely the weakest areas of health care-associated infection surveillance in China. Second, the hospitals participating in the surveillance in various provinces/municipality were not randomly selected, but were selected by voluntary registration. There is also no requirement on the number of participating hospitals and ICUs in each province, resulting in a small number of participating hospitals and ICUs in some provinces. Third, because there is no reporting software system, the accuracy of data cannot be actively controlled and screened, which increases the workload of manual reporting and review. Therefore, during the surveillance process, some hospitals withdrew from the surveillance group due to insufficient quality of data reporting or heavy reporting workload. Despite these limitations, our data represent the trend of CLABSI epidemiology in 2014-2018 for most tertiary general public hospitals in China.

Conclusion
The average incidence of CLABSI in this study was 1.5/1000 catheter days, lower than previous reports in China, but similar to the rate reported in developed countries. As a continuous surveillance project, targeted surveillance of ICU CLABSI cases can not only detect the changing trend of incidence and pathogens, but also find the regional differences in CLABSI incidence. It is necessary to implement targeted surveillance of CLABSI cases by using standardized CLABSI surveillance definitions and methodologies.