Journal of Hospital Infection 87 (2014) 1e10 Available online at www.sciencedirect.com

Journal of Hospital Infection journal homepage: www.elsevierhealth.com/journals/jhin

Review

Blood culture contaminants S. Dawson* Department of Microbiology, Great Western Hospital, Swindon, UK

A R T I C L E

I N F O

Article history: Received 3 October 2013 Accepted 23 February 2014 Available online 26 March 2014 Keywords: Blood cultures Contamination Cross-infection

S U M M A R Y

Blood cultures are an essential diagnostic tool. However, contamination may impact on patients’ care and lead to increased patient stay, additional tests, and inappropriate antibiotic use. The aim of this study was to review the literature for factors that influence the rate of blood culture contamination. A comprehensive literature search was performed using Medline and CINAHL on blood culture contamination. Hospitals/units should have in place a protocol for staff on how to take blood cultures, incorporating use of an aseptic technique. Studies have shown that several key factors in the process may lower contamination rates such as adherence to a protocol, sampling by peripheral venepuncture route rather than via an intravascular catheter, use of sterile gloves, cleaning tops of blood culture bottles with antiseptics and inoculating blood culture bottles before other blood tubes, samples being taken by a phlebotomy team, monitoring contamination rates, and providing individual feedback and retraining for those with contaminants. Although skin antisepsis is advocated there is still debate on which antiseptic is most effective, as there is no conclusive evidence, only that there is benefit from alcohol-containing preparations. In conclusion, hospitals should aim to minimize their blood culture contamination rates. They should monitor their rate regularly and aim for a rate of 3%. ª 2014 The Healthcare Infection Society. Published by Elsevier Ltd. All rights reserved.

Introduction Blood cultures are used in clinical practice to investigate a suspected systemic infection. Those that grow a true pathogen can assist clinicians with identifying the cause of the patients’ sepsis and provide antibiotic sensitivities, whereas growth of contaminants may be detrimental to patients’ care. Contamination results from a number of sources: the patient’s skin; the equipment used to take the sample and transfer it to the culture bottle; the hands of the person taking the blood sample; or the general environment.1 It can lead to clinical misinterpretation and result in extra expense due to increased length of stay of patients, additional laboratory tests, and use of inappropriate antibiotics.2e4 A recent retrospective matched caseecontrol study, which evaluated 142 false positives and * Address: Department of Microbiology, Great Western Hospital, Swindon SN3 6BB, UK. Tel.: þ44 (0) 1793 604800. E-mail address: [email protected].

controls, found a significant difference in means between cases and controls in relation to total costs [£5,001.5 (US$7,502.2); 95% confidence interval (CI): £3,283.9 ($4,925.8) to £6,719.1 ($10,078.6); P < 0.001] and length of hospital stay of 5.4 days (95% CI: 2.8e8.1; P < 0.001).2 Similarly Gander et al. found a median $8,720 additional charge for each contaminated event compared to negative cultures, but in their study the median length of stay only increased marginally from four days (95% CI: 4e5) to five days (95% CI: 4e7).4 Contaminants can also result in antibiotics being given unnecessarily.3,5,6 In a multivariate analysis contaminants were independently correlated with a 39% increase in intravenous antibiotic charges.3 In their 12-week study, Souvenir et al. found that antimicrobials were used to treat nearly one-half of patients with contaminated blood cultures, with vancomycin being misused in 34% of patients; Lee et al. also noted that 41% of their pseudobacteraemia cases were unnecessarily treated with systemic antibiotics, of which glycopeptides accounted for 20%.5,6 Algorithms have now been developed to assist

0195-6701/$ e see front matter ª 2014 The Healthcare Infection Society. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jhin.2014.02.009

2

S. Dawson / Journal of Hospital Infection 87 (2014) 1e10

clinicians with differentiating pathogens from contaminants in blood cultures. Weinstein and Doern consider that in addition to clinical findings (e.g. fever, leucocytosis) the two most valuable criteria are the identity of the organism and the number of blood culture sets positive vs number obtained.7 The time to detection of the positive result and number of bottles positive in a set are not of value, as there is too much overlap between true pathogens and contaminants.7 The development of new, rapid, organism identification techniques such as matrix-assisted laser desorptioneionization time-of-flight mass spectrometry and polymerase chain reaction will enable laboratories to speed up the identification of organisms from blood samples and cultures, thereby limiting misinterpretation, over-investigation and inappropriate treatment for contaminants.8,9

Methods A comprehensive literature search was performed, which included during December 2013 using Medline and CINAHL to retrieve papers in English from 1990 onwards using the search terms ‘blood culture contaminant’ and ‘blood culture contamination’.

Contamination rates Variation in rates between hospitals Reports from NHS Trusts and equipment suppliers suggest that rates could be as high as 10%.10 However, it is accepted that the complete elimination of all blood culture contamination is not feasible, especially in patients where venous access is difficult and aseptic technique is compromised.11 The Clinical and Laboratory Standards Institute recommends that an acceptable range should be 3%.12 Similarly the Department of Health (England) has suggested that contamination rates should be 0.9)

Intervention: insertion of intravascular catheter samples in paediatric population Ramsook et al.20 Paediatric emergency Single aerobic BC bottle taken from department 2431 patients either via IV catheter or peripheral venepuncture. Children’s hospital Pre-intervention cultures were Norberg et al.21 emergency department obtained simultaneously with PIV catheter insertion and postintervention specimens were taken by a separate dedicated procedure Paediatric emergency Phlebotomy policy was changed so Weddle et al.36 department that BCs had to be obtained by a second venepuncture and not during insertion of PIV catheter.

Isaacman et al.37

Paediatric emergency department

Hall et al.38

Paediatric emergency department

Intervention: antisepsis of bottle tops Schifman et al.22 640 institutions

BC CR for samples taken by IV catheter: 3.4% (44/1295) compared to 2% (22/ 1084) for venepuncture (P ¼ 0.043) 4108 BCs obtained (pre intervention: 2108; post intervention: 2000). BC CR decreased from pre 9.1% to post 2.8% (P < 0.001).

Pre-intervention BC CR: 6.7% (120/ 1796) Post-intervention BC CR: 2.3% (23/ 1229) Significant difference: P ¼ 0.001 OR: 2.96 (CI: 1.96e4.57) BCs obtained in 99 children through CR for both methods 1.0% newly inserted IV catheter and (95% CI: 0e3.0%) butterfly needle at separate venepuncture site Introduction of a sterile BC BC CR reduced from 3.9% to 1.6% during collection process via PIV catheter intervention period (P < 0.001) and development of web-based education Q-Probes study of the College of American Pathologists

Intervention: type of gloves Kim et al.45 Single-centre trial 10,520 BCs taken by interns in a involving medical wards prospective cluster randomized and intensive care unit cross-over controlled trial. Interventions were use of sterile gloves every time venepuncture performed or optional sterile gloving (worn only if needed such as repalpating vein after disinfection)

95.5% of laboratories routinely applied antiseptic to the top of the broth bottle before specimen inoculation. This group had a significantly lower CR (2.3%) compared to the remaining laboratories (3.4%) (P ¼ 0.018). BC classified as possible contaminants. CR was 0.6% for routine sterile gloving and 1.1% for optional (adjusted OR: 0.57; 95% CI: 0.37e0.87; P ¼ 0.009). BC classified as likely contaminants. CR was 0.5% for routine sterile gloving and 0.9% for optional (adjusted OR: 0.51; 95% CI: 0.31e0.83; P ¼ 0.007). (continued on next page)

4

S. Dawson / Journal of Hospital Infection 87 (2014) 1e10

Table I (continued ) Study

Site 48

Self et al.

Emergency department

Intervention: needle changes Meta-analysis Spitalnic et al.49

Intervention: use of prepacked kits/packs Thomas et al.11 UK hospital Snyder et al.29

Meta-analysis

Bamber et al.54

UK hospital

Dhillon et al.55

UK hospital

Weightman and Kerr56 UK hospital

Marini et al.57

Emergency department

Details

Change of technique from Baseline period BC CR: 4.3% traditional clean procedure to new Intervention period BC CR: 1.7% sterile procedure with use of sterile (P < 0.001) gloves, new materials kit containing 2% chlorhexidine skin antisepsis device, sterile fenestrated drape, and a procedural checklist. Monitored using an interrupted timeseries study. Analysis of eight studies that compared BC CR with or without a needle change prior to inoculation of BC bottles.

356 institutions

Weighted contaminant rate of 2% if needle changed prior to inoculation compared to 3.7% if needle not changed. Weighted overall difference: 1.25% (95% CI: 0.75e1.75%; P < 0.001).

Introduction of BC packs and training for staff using kits. Seven studies comparing BC CR when pre-packaged kit was used for taking BCs

BC CR reduced from 9.2% to 3.8% following pack introduction. Not statistically significant. Mean OR: 1.12; 95% CI: 0.94e1.35. No recommendation made for or against using pre-packaged kits. Introduction of BC packs and staff BC CR reduced from 43% to 25% of total training number of positive BC. Introduced BC packs and training BC CR significantly reduced from 8.7% programme to 3%. P < 0.001 following introduction of packs. Introduction of BC kits and staff BC CR reduced from 6% prior to training. In addition BC CRs were fed interventions to an average of 2.7% post back to each clinical directorate interventions every three months. Introduction of BC packs with BC CRs reduced from 2.1% to 1.4% with education for staff. Feedback on CR intervention provided.

Intervention: which type of healthcare worker takes BC sample Gander et al.4 Emergency department Total of 5432 BCs were taken from of 968-bed tertiary care two emergency departments (EDs): teaching hospital ED west and ED non-west.

Bekeris et al.13

Results

Q-Tracks study of the College of American Pathologists

ED west: BC CR Phlebotomy: 3.1% Non-phlebotomy: 7.4% Significant difference: P < 0.001 ED non-west: BC CR Non-phlebotomists: 5.6% BC CR phlebotomists ED west and nonphlebotomists ED non-west significant difference: P < 0.001 BC CR significantly lower in institutions using dedicated phlebotomy team (P < 0.001). Institutions where nursing staff did not collect BCs had average BC CR of 2.17% and those where virtually all BC collected by nursing personnel had average BC CR of 4.21%.

S. Dawson / Journal of Hospital Infection 87 (2014) 1e10

5

Table I (continued ) Study

Site 20

Ramsook et al.

Paediatric emergency department

Schifman et al.22

640 institutions

Snyder et al.29

Meta-analysis

Weightman and Kerr56 UK hospital Weinbaum et al.59

Medical and surgical units in an acute-care community hospital with 487 beds.

Surdelscu et al.60

Community teaching hospital

Details

Results

Single aerobic BC taken from 2431 patients either via IV catheter or peripheral venepuncture. Nurses and laboratory phlebotomists used different antiseptics protocols for skin preparation

BC CR: venepuncture, IV catheter, overall (nurse: 1.2%, 3.4%, 2.8%; phlebotomist: 2.6%, e, 2.6%, respectively). Overall CR did not significantly differ for phlebotomists and nurses 2.6% vs 2.8% (P ¼ 0.807) but for venepuncture only nurses had a lower CR. Q probes study of the College of CR significantly higher (2.7%) in American Pathologists institutions if

Blood culture contaminants.

Blood cultures are an essential diagnostic tool. However, contamination may impact on patients' care and lead to increased patient stay, additional te...
296KB Sizes 0 Downloads 3 Views