Incidence of catheter-related bloodstream infections

Central intravascular catheters (CVC) are essential in the treatment of intensive care patients (ICU), who require dialysis, parenteral medications or nutrition, or those under haemodynamic monitoring.^1–3 According to the recommendation of the Centre for Disease Control and Prevention (CDC) in 1992, needleless connector devices were designed and introduced, with the goal of reducing the risk of sharps injuries among healthcare personnel, as well as improving patient healthcare safety.^4-6

Although needleless connectors have fulfilled these functions, several challenges were associated with their introduction, such as catheter-related bloodstream infections (CRBSIs) and catheter lumen occlusion.^5–11 Acquisition of CRBSI is linked to an increased financial burden in terms of high mortality rate, extended in-hospital stay, prolonged duration of supportive ventilation and an excessive medications and other medical consumable materials expenditures.^{3,5–6,11–12} A total of 200,000 central venous catheters (CVC) were used annually in the UK in 2001.¹ The Health Protection Agency (UK) reported that blood stream infections increased in the UK from 80,000 in 2003 to 105,000 in 2007.⁹ Of these, 44% were deemed to be related to IV access devices.¹⁰

The majority of needleless connectors are classified into two groups, which are different structurally and functionally: simple split-septum needleless connectors or complex (have internal mobile components) mechanical-valve connectors.^5–6 Despite the report of the CDC (2002),¹³ in which it was demonstrated that the proper use of needleless connectors, as per the manufacturer recommendations and instructions, does not have a significant negative impact in terms of the incidence rate of healthcare acquired catheter-related bloodstream infection (CRBSI), there have been contradictory and conflicting opinions and findings regarding the actual impact of introducing these devices.^1,5–6,8,14 As per the findings of many investigators, the sophisticated design of MV luer access devices was deemed the most significant reason for difficulty in disinfection of these devices.

The internal engineering complicates the cleaning process for the area that lies between the inner parts of the connector and the catheter hub.^{5–6,15–16} In addition, the 2011 guidelines of CDC recommended using split-septum needleless connectors rather than utilising sophisticated mechanical-valve connectors.^6,14 However, there has been a dispute as whether a greater risk of CRBSIs was linked to certain type of needleless connectors. The objectives of this study were: (i) to assess any change in CRBSI incidence rates when replacing mechanical-valve needleless connector with split-septum connectors; and (ii) to identify risk factors correlated with CRBSI within the studied ICU.

The study was implemented in the ICU in Antrim Area Hospital in Northern Ireland, a 426-bed general teaching hospital serving a population of approximately 420,200. Acute general medical and surgical services are provided by this hospital. Additionally, many primary outpatient healthcare facilities are supported by this hospital, which is considered an important centre for the coordination of health services across a defined geographical area in Northern Ireland. This project was approved by the Northern Health and Social Care Trust (NHSCT) research governance office as a service development project, and it was recorded on the hospital’s Audit and Effectiveness Quality Improvement Plan.

Central venous catheter–related bloodstream infection incidence (CRBSI): The Centres for Disease Control and Prevention (CDC) criteria, in identifying a CRBSI, were applied as follows: (i) an identified pathogen (isolate) sampled from one or more blood cultures that was not relevant to an infection(s) at another body site(s); (ii) the patient had a CVC at the time or within 48 hours before the culture was obtained; (iii) signs and symptoms of a bloodstream infection (for example, chills, fever or hypotension), or positive blood culture results generated by the microbiology laboratories, were not connected to an another body site(s) infection (for example, it was mainly of unknown origin); and (iv) common skin contaminants were cultured from two or more blood cultures drawn, on separate occasions, from any site [that is, Diphtheroid (Corynebacterium), Bacillus, Streptococcus viridians, coagulase-negative staphylococci].

Baseline period: The baseline period was defined as the period when only mechanical-valve needleless connector devices (MV-NC; Clave®, ICU medical, USA) were exclusively used in the ICU.

Post-intervention period: The post-intervention period was defined as the period when split-septum needleless connectors (SS-NC; BD Q- SYTE®) were exclusively used in the ICU.

Central venous catheter-day (CVC-day): A CVC-day represents a single day on which one or more central lines were maintained in place for an individual patient.

In order to determine the CRBSI rate, the number of CRBSIs (through determining the true-positive blood cultures resulting from central venous catheter use), and the count of bed-days for all patients admitted to the ICU department, over the study period, were documented on a monthly basis. This was achieved by reviewing the ICU department database, as well as the microbiology laboratory records (retrospectively during the period from July 2013 to December 2013 and prospectively during the period from January 2014 to June 2014).

The required data were gathered using a bespoke data collection form, which allowed recording of: patient’s date of birth, gender, Acute Physiology and Chronic Health Evaluation II score (APACHE II, which reflects the patient’s severity of illness), infecting organism(s) type, central line(s) type and location, date and time of catheter(s) insertion, presence of previous bloodstream infection(s), immunodeficiency (especially neutropenia), receipt of parenteral nutrition (PN), medical specialty, the patient location prior to admission to the ICU, the catheterisation period, whether the patient was receiving dialysis (through central venous catheter) and the patient diagnosis(es).

This observational (before/after) study design was used. All patients were tracked from their admission to the ICU until discharge over the complete study period. During the first stage, a retrospective review of the hospital ICU department database and clinical microbiology laboratory information system was conducted. This review was carried out to measure the monthly rate of CRBSI for the period from July 2013 to December 2013 (control period).  Blood culture results, obtained from samples taken in the first 48 hours of admission, were not considered, that is, only hospital acquired infections were considered. Data were also collected retrospectively during this control period in an attempt to identify risk factors for CRBSI.

In the second stage of the study, the new device was introduced into the ICU and similar data were collected on a monthly basis for the period January 2014 to June 2014 (intervention period). Education on the proper use and disinfection of the SS-NC device was provided to the healthcare personnel in the ICU, by nurse educators from Antrim Area Hospital. During the overall period of the study, CRBSI prevention practices were not changed, that is, hand hygiene practices for healthcare staff, use of suitable patient skin antiseptic for CVC placement and barrier precautions for the CVC connection taking account of the relevant guideline recommendations (that is, to reduce CVC contamination hazards).

The primary outcome measure was the rate of CRBSI in hospitalised patients during the intervention period (split-septum needleless connectors in place), compared with the control period. The secondary outcome measure related to risk factors associated with the development of CRBSI.

Descriptive statistics were employed to describe both patient demographics and other baseline characteristics. CRBSI rates over both study periods were compared using statistical run charts and the Pearson Chi Square test. To evaluate the association between continuous independent variables (age, APACHE II score, catheterisation period, or the length of stay in the ICU) either the Student t-test or Mann-Whitney U-test was used, depending on the normality of data distribution. The Pearson Chi Square test was used with categorical variables (that is, age categories, previous bloodstream infection, neutropenia, etc.).

To identify independent risk factors for CRBSI, all variables that showed significant associations (with p value ≤0.25) with CRBSI on univariate analysis, were entered into a backward multiple variable logistic regression model. A P value ≤0.05 was deemed significant in logistic regression modelling.  All statistical analyses were carried out using the Statistical Package for Social Sciences (Windows Version 20.0; SPSS®, Chicago).

Throughout the duration of the study (baseline and post-intervention periods), a total of 380 patients were admitted to the ICU. Of these, 197 patients were admitted during the intervention period and 183 patients were admitted during the control/baseline period. The demographics and baseline characteristics of these patients are presented in Table 1. Aspects of the central venous catheters inserted over both study phases (CVC body site(s) and combined total catheter-days) are also shown in Table 1. These latter data showed that there were no significant differences with regard to several variables; however, the average APACHE II score was higher (16.3 versus 14.4; p=0.005) in patients included in the intervention period when compared to baseline patients (14.4; p=0.005).

Of the 380 patients included in the study, nine patients were diagnosed with CRBSI prior to commencing the intervention while five patients were diagnosed with CRBSI during the period after the introduction of the new connector device. The incidence density of CRBSI revealed 12.20 cases/1000 CVC-days (that is, equivalent to 0.80 cases/100 bed-days) during pre-intervention period, whereas it was 7.35 cases/1000 CVC-days (that is, equivalent to 0.51 cases/100 bed-days) after the intervention. Moreover, the estimated percentage of patients who developed CRBSI was 4.6% prior to the intervention and 2.7% after the intervention period. The intervention was associated with a reduction in the CRBSI incidence rate by 41.3%; however, this was not statistically significant (p=0.342).

To compare the change in incidence rates during both phases of the study, statistical run charts were plotted as shown in Figures 1 and 2. The run chart (Figure 1) shows that the mean line slope dropped from 14.03 cases/1000 CVC-days, pre-intervention, to reach 5.80 cases/1000 CVC-days, after the intervention was put in place. Furthermore, as indicated in Figure 2, the slope of the mean line declined sharply from 0.90 cases/100 bed-days to 0.45 cases/100 bed-days, between the baseline and post-intervention periods. There was an increase in slope of trendline plotted during the baseline period (Figures 1, 2). These latter findings help define the effectiveness of the new device.

Figure 1. Run-chart showing CRBSI incidence rates (normalised per 1000 CVC-days) plotted at monthly intervals, over the entire study period (July 2013 – June 2014). The vertical dashed line represents the date of the introduction of split-septum connectors into the ICU. The red dashed line refers to trendline

Figure 2 Run-chart showing CRBSI incidence rates (normalised per 100 bed-days) plotted at monthly intervals, over the entire study period (July 2013 – June 2014). The vertical dashed line represents the date of introduction of split-septum connectors into the ICU. The red dashed line refers to the trendline

A variety of pathogens were detected during the baseline and post-intervention periods. These organisms included gram-positive bacteria (for example, Enterobacter cloacae and Staphylococcus aureus), gram-negative bacteria (for example, Escherichia coli and Pseudomonas aeruginosa) and fungi (for example, Candida albicans and Candida parapsilosis; Table 2).

The univariate analysis (Table 3) of several variables related to CRBSIs demonstrated a trend of the following variables to be higher (with p values ≤0.25) among patients diagnosed with CRBSI: presence of previous bloodstream infection (p=0.014), presence of neutropenia (p=0.008), CVC body site-femoral (p=0.009), CVC body site-jugular (p=0.009), body site-subclavian (p=0.096), male gender (p=0.071), receiving parental nutrition (p=0.084), receipt of dialysis (p<0.001). Furthermore, the average APACHE II score was greater within the CRBSI patient group, compared to the non-CRBSI group (p=0.018). In addition, the mean length of stay in the ICU (p=0.004) and the mean catheterisation period (catheter-days; p<0.001) were longer for patients with CRBSI, in comparison to those without CRBSI. The demographics and baseline characteristics of patients within the CRBSI and non-CRBSI groups are displayed in Table 3.

Carrying out a backward multivariable logistic regression analysis using variables that showed an association with CRBSI (that is, with p value ≤0.25) indicated the following variables as significantly independent predictors for CRBSI: having neutropenia (coefficient= 2.145, p=0.022; OR, 8.54, 95% CI: 1.36 – 53.78), and receiving dialysis (coefficient = 1.998, P=0.001; OR, 7.37, 95% CI: 2.25 – 24.17). This multivariate analysis model showed a good ability to predict the occurrence of CRBSI during the studied period, that is, the relevant percentage of correct prediction was found to be 94.9% (Table 4).

An increased rate of CRBSI within healthcare settings was associated with the introduction of needleless connectors in 1990.^6,11 The acquisition of CRBSI has serious implications on patient outcomes including an excess mortality rate, higher morbidity rate and increased attributable healthcare costs.^3,5,12,17

Baseline characteristics (for example, age, age group, gender, diagnosis, medical specialty, source of admission to the ICU, length of stay in the ICU, catheterisation period, existence of neutropenia, parenteral nutrition administration and dialysis reception, presence of former bloodstream infection, CVC jugular and CVC femoral), of the ICU patients did not differ significantly (P >0.05) before and after introducing the split-septum device. This stability in patient profile helped ensure that any change in the outcome of interest (CRBSI incidence rate) was caused solely by introducing the split-septum connectors (the intervention) rather than by other confounding factors.

The estimated average APACHE II scores in the post-intervention patient group were, however, greater than those estimated for baseline patients. This latter difference in APACHE II scores indicated that the post-intervention patients were ‘sicker’ than pre-intervention patients, which, in turn, strengthens the validity of the present findings, especially since an increased APACHE II score was associated in the univariate analysis, with an increased risk of CRBSI.

The present findings showed a decline in the CRBSI incidence rate after the introduction of the split-septum luer access devise into the ICU study site. The CRBSI incidence rate decreased considerably from 12.20 to 7.35 cases per 1000 CVC-days, amounting to a 41.3% reduction, post-intervention. However, the reduction was not statistically significant (that is, the small number of cases identified over the post-intervention period impacted on the estimated significance level).

A central venous care bundle was applied in the ICU of Antrim Area Hospital during the full study period, to ensure optimal CVC insertion and to sustain appropriate maintenance, that is, including applying hand hygiene practices, using a chlorohexidine based product for skin disinfection, applying barrier precautions for CVC placement, using alcohol for needleless connector cleaning. The compliance rate to this CVC care bundle was very high (approximately 99.0%) as per the annual Healthcare-Acquired Infection (HCAI) Performance Indicator Report generated by the infection control team within Northern Health and Social Care Trust.

The present study showed that CRBSIs, identified during the post-intervention period (split-septum device in place), were caused only by gram-negative bacteria and fungi. This finding is in line with findings of Salgado et al. (2007) who reported that a high percentage of CRBSIs diagnosed were due to gram-negative bacteria.¹⁶ Achievable APACHE II scores range from 1 to 71.^18–19 In the present study, the average APACHE II score values were much higher in the patients diagnosed with CRBSIs (19.3), when compared with patients who did not develop CRBSIs (on average 15.4).

This scoring index presents classification for a severity of illness, that is, a patient with higher APACHE II score is a sicker patient. Severely ill patients (within the ICUs) frequently receive intravenous medicines, intravenous fluids and parenteral nutrition through CVC. Furthermore, patients with greater APACHE II scores are often immune-compromised, as they receive immune-suppressive treatment (for example, cancer patients), and they are probably more exposed to a higher extent of radiation compared with those with lower APACHE II scores.¹¹

In the present study, the occurrence of neutropenia (for example, due to administration of chemotherapy in cancer patients) was found to be independently associated with an increased risk of CRBSI development. When the host’s immune system is weakened, the susceptibility of the host to infection is increased. For instance, the endogenous normal flora that are already established within a host’s own body (respiratory tract system, genitourinary tract system and skin surface), may become pathogenic. Staphylococci species (representing a major component of skin flora) have been reported as the most common causative organisms for CRBSI after undergoing central venous catheter insertion.^20–24

The receipt of dialysis through central venous catheters was found to be an independent predictor for the acquisition of bloodstream infections in our patients.

This infection primarily originates from the biofilm, which exists on the internal surface of the central line (the bacteria, which form this biofilm, are usually attained due to the repeated handling of the central line hub). Following this, the microbes move on from the intraluminal biofilm into the blood circulation (that is, bacterial seeding) and cause bloodstream infection (BSI). The mechanism illustrating microbial seeding is still not, fully understood, however, it may rely upon the type of infecting pathogen.^25–27

This study has a number of strengths. Awareness sessions on the appropriate use of needleless connector devices (including the optimal methods and techniques regarding appropriate disinfection, flushing, clamping, replacing, cleaning and locking of the device) were given to the ICU healthcare providers, over the period that preceded the introduction of the new connector. The study researcher and the ICU nurse educator participated in delivering these educational activities. In addition, the researcher and the ICU nurse educator provided advice to the ICU staff on any connector-related issues, during the entire intervention period.  Furthermore, the CRBSI incidence rate was normalised per 1000 CVC-days and were also expressed per 100 bed-days, which allowed comparison of the present findings with previous literature findings and other benchmark data.

The main limitation of the research was the fact that this was a single centre investigation within an ICU unit with limited bed-occupancy. The internal validity of the present design and the generalisability would have been improved if the study had occurred within multiple intensive care units. Secondly, although a reasonable sample size of patients (a total of 380) was included over the two study periods, the intervention period was relatively short (six months). Thirdly, the compliance rate to the facility guidelines (regarding infection control practices and the appropriate use of the connector devices), by the ICU staff, was not surveyed over the two study periods. Nonetheless, the compliance rate to CVC care bundle within the ICU during the study period was very high (approximately 99% as per the annual Healthcare-Acquired Infection (HCAI) Performance Indicator Report generated by the infection control team within Northern Health and Social Care Trust.

In conclusion, the introduction of a split-septum luer access device led to a decrease in the CRBSI incidence rate in the ICU study site, although the results did not reach statistical significance. The present study benchmarked the CRBSI incidence rate in the ICU at the Antrim Area Hospital (to the best of our knowledge, the present investigation is the first piece of work of its kind in the United Kingdom). Neutropenia and undergoing dialysis were determined as potential risk factors for acquisition of CRBSI.

• The purpose of this study was to evaluate any change in the incidence of CRBSI when switching from mechanical-valve connectors to split-septum needleless connectors.
• This study was observational (before/after) in design, and a total of 380 patients were included in this study throughout both the baseline and intervention periods.
• A variety of pathogens were detected during the baseline and post-intervention periods. These organisms included gram-positive bacteria, gram-negative bacteria and fungi.
• The introduction of split-septum connectors contributed to a trend toward a reduction in the CRBSI incidence rate.
• Having neutropenia and being on dialysis were statistically and significantly associated with detection of CRBSI.

1. Casey A, Spare M, Worthington T. Letters to the editor. J Hosp Infect 2002;50:77–81.
2. Maragakis LL, Bradley KL, Song, X. Increased catheter-related bloodstream infection rates after the introduction of a new mechanical valve intravenous access port. Infect Control Hosp Epidemiol 2006;27:67–70.
3. Touré A et al. Propensity score analysis confirms the independent effect of parenteral nutrition on the risk of central venous catheter-related bloodstream infection in ontological patients. Clin Nutr 2013;32:1050–4.
4. Jarvis WR et al. Health care-associated bloodstream infections associated with negative or positive-pressure or displacement mechanical valve needleless connectors. Clin Infect Dis 2009;49:1821–7.
5. Btaiche IF et al. The effect of needleless connectors on catheter-related bloodstream infections. Am J Infect Control 2011;39:277–83.
6. Hadaway L. Needleless Connectors for IV Catheters. Am J Nurs 2012;112:32–44.
7. Field K et al. Incidence of catheter-related bloodstream infection among patient with a needleless, mechanical valve-based intravenous connector in an Australian haematology-oncology unit. Infect Control Hosp Epidemiol 2007;28:610–3.
8. Rupp MA et al. Outbreak of bloodstream infection temporally associated with use of an intravascular needleless valve. Clin Infect Dis 2007;44:1408–14.
9. Health Protection Agency (2008) Surveillance of healthcare associated infection report 2008. http://www.isb.nhs.uk/library/standard/178 (accessed March 2013).
10. National Audit Office (NAO; 2009) Reducing Healthcare associated infection in England. www.nao.org.uk/wp-content/uploads/2009/06/0809560es.pdf (accessed October 2013).
11. Wittekamp BH et al. Catheter-related bloodstream infections: a prospective observational study of central venous and arterial catheters. Scand J Infect Dis 2013;45:738–45.
12. Rosenthal VD. Central line–associated bloodstream infections in limited‐resource countries: A review of the literature. Clin Infect Dis 2009;49:1899–907.
13. Centres for Disease Control and Prevention (2002) Recommendations and Reports, 51(RR10), 27–28. http://www.cdc.gov/mmwr/preview/mmwrhtml/rr5110a2.htm (accessed March 2013).
14. O’Grady NP, Alexander M, Dellinger EP. (2002) Guidelines for the prevention of intravascular catheter-related infections. Centres for Disease Control and Prevention. MMWR Report, 51, 1–29.http://www.cdc.gov/mmwr/preview/mmwrhtml/rr5110a2.htm (accessed March 2013).
15. Menhay SZ, Maki DG. Disinfection of needleless catheter connectors and access ports with alcohol may not prevent microbial entry: the promise of a novel antiseptic barrier cap. Infect Control Hosp Epidemiol 2006;27:23–7.
16. Salgado CA et al. Increased rate of catheter-related bloodstream infection associated with use of mechanical valve devise at a long-term acute hospital. Infect Control Hosp Epidemiol 2007;44:1408–14.
17. Daniels KR, Frei CR. The United States’ progress toward eliminating catheter-related bloodstream infections: Incidence, mortality, and hospital length of stay from 1996 to 2008. Am J Infect Control 2013;41:118–21.
18. Edwards JR, Paterson KD, Andrus ML. National Healthcare Safety Network (NHSN) Report, data summary for 2006, issued June 2007. Am J Infect 2007;35:290–301.
19. Knaus WA et al. APACHE II: A severity of disease classification system. Crit Care Med 1985;13:818–29.
20. Ducel G, Fabry J, Nicolle L. (2002) Prevention of hospital-acquired infections: A practical guide. Geneva: The World Health Organization. www.who.int/csr/resources/publications/whocdscsreph200212.pdf (accessed September 2014).
21. Crogan NL, Evans BC. Clostridium difficile: An emerging epidemic in nursing homes. Geriatr Nurs 2007;28:161–4.
22. Carroll KC, Bartlett JG. Biology of Clostridium difficile: Implications for epidemiology and diagnosis. Ann Rev Microbiol 2011;65:501–21.
23. Raad I et al. Sources and outcome of bloodstream infections in cancer patients: the role of central venous catheters, Eur J Clin Microbiol Infect Dis 2011;26:549–56.
24. Wojewoda CM et al. Bloodstream infection caused by non-toxigenic Corynebacterium diphtheria in an immune-compromised host in the United States. J Clin Microbiol 2012;2170–2.
25. Betjes MGH. Prevention of catheter-related bloodstream infection in patients on haemodialysis, Nature Rev Nephrol 2011;7:257–65.
26. Camins BC. Prevention and treatment of haemodialysis-related bloodstream infections. Sem Dialys 2013;26:476–81.
27. Abebe A et al. Catheter-related bloodstream infections reviews. Clin Key 2014;3:e32–e49