Diagnostic Stewardship: A Systematic Review and Meta-analysis of Blood Collection Diversion Devices Used to Reduce Blood Culture Contamination and Improve the Accuracy of Diagnosis in Clinical Settings

Abstract Background Blood culture contamination may lead to misdiagnosis, overutilization of antibiotics, and prolonged length of stay. Blood specimen diversion devices can reduce contamination rates during blood culture collection procedures. We performed a systematic literature review and meta-analysis evaluating the influence of blood specimen diversion devices in blood culture contamination rates. Methods We searched Medline, Cumulative Index to Nursing and Allied Health Literature, Embase, Cochrane, Scopus, and Web of Science, from database inception to 1 March 2023, for studies evaluating the impact of a diversion device on blood culture contamination. Blood culture contamination was a positive blood culture with microorganisms not representative of true bacteremia, but rather introduced during collection or processing the blood sample. Random-effects models were used to obtain pooled mean differences, and heterogeneity was assessed using the I2 test. Results Of 1768 screened studies, 12 met inclusion criteria for this systematic literature review. Of them, 9 studies were included in the meta-analysis. Studies were substantially heterogeneous, but stratified analyses considering only high-quality studies revealed that venipuncture using a diversion device was associated with a significant reduction in blood culture contamination in comparison to the standard procedure of collection (pooled odds ratio [OR], 0.26 [95% confidence interval {CI}, .13–.54]; I2 = 19%). Furthermore, the stratified analysis showed that the adoption of a diversion device did not reduce the detection of true infection (pooled OR, 0.85 [95% CI, .65–1.11]; I2 = 0%). Conclusions Blood culture diversion devices was associated with decreased contamination rates and could improve quality of care, reduce costs, and avoid unnecessary antibiotic use.

Around 30 million blood cultures are routinely collected annually in the United States for the diagnosis of bacteremia [1].Most of these hospitals report a contamination rate of 2%-3% [1].However, 20%-50% of all positive cultures are likely to be falsely positive [2,3], leading to unnecessary broadspectrum antibiotic therapy, prolonged hospital stay, and additional hospital costs [1,4].The American Society for Microbiology recommends a performance standard rate of <3% for blood culture contamination in medical institutions [5], but recent studies suggest that lower contamination rates may be achieved through the implementation of diagnostic stewardship elements [6].In the last practical guideline for clinical microbiology laboratories, the Centers for Disease Control and Prevention registered its institutional recommendation of considering ≤1% as a new universal standard for allowable contamination rates, given that many facilities have already achieved this proportion following best practices [7,8].
While standard procedures such as skin antisepsis and bottle disinfection have shown positive outcomes in reducing blood culture contamination rates [9], some studies show that further reduction may be achieved by discarding the initial blood draw portion, which potentially contains skin microorganisms [10][11][12].In a recent review, 2 main procedures of diversion were presented: the open technique and the use of a diversion device [13,14].
The implementation of a diversion device to remove potentially contaminated initial blood samples is being tested in some health institutions [13].This idea converges with the concept of diagnostic stewardship, which aims to optimize diagnostic testing, thus reducing diagnostic error and consequent unneeded therapy [15].This systematic literature review and meta-analysis aims to investigate the efficacy of initial blood specimen diversion devices in reducing blood culture contamination rates.

Systematic Review and Search Strategies
This systematic literature review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement [16] and the Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines [17].This study was registered on the International Prospective Register of Systematic Reviews (PROSPERO; https://www.crd.york.ac.uk/PROSPERO/) on 20 March 2023 (CRD42023404440).

Search Strategy
Our search strategy was developed in March 2023 with the assistance of a health sciences librarian with expertise in systematic reviews.We explored Medline (PubMed), Cumulative Index to Nursing and Allied Health Literature (CINAHL), Cochrane Central, Web of Science, and Scopus and Embase (Elsevier platform).The literature search included publications from database inception to 1 March 2023.The full PubMed search strategy (Supplementary Table 1) was adapted for the other databases.We searched only English-language literature.
Studies were included if they evaluated the impact of a blood diversion device on blood culture contamination.Blood culture contamination was defined as the presence of microorganisms in a blood culture that were not representative of a true bloodstream infection, but rather a false-positive result during collection or processing of the blood sample [4].We excluded comments or reviews, studies without a control group, pilot studies, studies of devices not integrated into the needle or catheter, studies using an open diversion technique, or those solely performed in children.
To filter the 1768 articles obtained from the databases, all titles and/or abstracts were examined (G.Y. C.), and those deemed unsuitable were excluded.After this first evaluation, all of the remaining articles were fully read, and 12 of them met the inclusion criteria and were included in the systematic review (Figure 1).

Data Abstraction and Quality Assessment
Of 5 independent reviewers (G.Y. C., V. L., E. T., T. A. M., and A. R. M.), 2 abstracted data for each article using a standardized  abstraction form (Supplementary Form 1).We recorded data regarding publication year, study period, design, population selection, setting, type of device, antisepsis procedure, blood culture contamination definition, outcomes, and financial analysis.We also collected information about the number of milliliters diverted, if they contained data on true bacteremia, and adjustment for potential confounders.Our primary outcome was the frequency of blood culture contamination in cultures collected with and without a diversion device.We used the Downs and Black scale [18] to evaluate study quality.All questions of the original published scale were answered for each reviewed paper and the total score calculated.We adapted question 27 of the Downs and Black scale replacing the multiple-choices options with a yes/no answer.The maximum score was 28.The reviewers performed component quality analysis independently, and inconsistencies were resolved by discussion.

Patient Consent Statement
The present investigation is a systematic literature review and meta-analysis of published data, so no patient informed consent was required.

Statistical Analysis
To perform the meta-analysis with the extracted data, blood cultures collected with a standard procedure and through a diversion device were assessed using a random-effects model to estimate pooled odds ratios (ORs) and 95% confidence intervals (CIs) with weights as described by DerSimonian and Laird [19].We performed stratified analyses considering study design, period of publication, duration, setting, tested device, detection system, collectors' profession, number of participants in each group, antisepsis technique, study quality, and others (Supplementary Table 2).We excluded studies not registering the total number of blood cultures collected.
Heterogeneity between studies was evaluated using the I 2 statistic the Cochran Q statistic test.We used the Cochrane Review Manager (RevMan) Web edition 4.12.0.Publication bias was assessed using funnel plot with RevMan (Supplementary Figure 2) and the Egger test with Comprehensive Meta-Analysis version 4 software (Biostat Inc, Englewood, New Jersey).
Of the 9 studies that provided a definition of blood culture contamination, only 4 explicitly mentioned the required number of bottles for classifying contamination [20,21,27,29].Among these, 3 studies considered the presence of microorganisms in either an anaerobic or aerobic bottle, but not both, as an indication of blood culture contamination [20,21,27,29].The other study considered contamination when microorganisms were detected in 1 of 4 bottles for matched sets or 1 of 2 bottles in both subsets [20,21,27,29].
Only one-quarter of the studies reported antibiotic use/duration, hospital length of stay (LOS), or mortality [23][24][25]31].The most reported outcome was LOS, in 3 studies [23,24,31].One study reported no difference in LOS between diversion device and standard procedures (P = .7)[23].Another 2 studies found significant reduction in LOS in the diversion device group (24 hours, P = .02)[31] and in the true-negative blood culture group (56 hours, P = .008)[24].Two studies evaluated diversion device impact on antibiotic use [23,25].One of them did not find a statistically different antibiotic duration (P = .19)[23], whereas the other found a 31.4% reduction in antibiotic use after diversion device implementation [25].The study reporting mortality data found a lower mortality in the diversion device group (P = .03)[31].
Overall, 9 studies, encompassing 83 325 collected blood cultures, evaluated contamination between those who used a diversion device and those who followed standard procedure of venipuncture (Supplementary Table 1 and Supplementary Figure 1) and were included in the meta-analysis [20,21,23,25,[27][28][29][30][31].Three studies that were included in the systematic review were not included in the meta-analysis because they did not report the total number of blood cultures [22,24,26].Upon analysis of the 9 studies, the degree of heterogeneity was high (P < .001;I 2 = 81%) (Supplementary Figure 1).However, even when considering 95% CIs, none of these studies demonstrated an advantage of using the standard procedure.
To mitigate this high level of heterogeneity, we performed a stratified analysis that took into account studies that were based on supplementary parameters (Supplementary Table 2 and Supplementary Figure 1).
When analyzing the 3 studies that were considered high quality by Downs and Black evaluation (≥18 points) and reported the total number of blood cultures [28,30,31], using a stratified analysis, we observed that diversion device adoption was associated with a reduced rate of blood culture contamination (pooled OR, 0.26 [95% CI, .13-.54]) with low heterogeneity results for the studies (P = .29;I 2 = 19%) (Figure 2).Conversely, the stratified analysis assessing true infection (3 studies) [28,30,31] showed no significant difference (pooled OR, 0.85 [95% CI, .65-1.11]), with homogenous results (P = .62;I 2 = 0%) (Figure 3).The potential for publication bias was assessed using 2 methods.The funnel plot (Supplementary Figure 2) appeared symmetrical and the Egger test did not indicate publication bias among the 9 studies included in the meta-analysis (P = .12).When the 3 high-quality studies were assessed by the Egger test, no publication bias was shown either (P = .13).

DISCUSSION
In this systematic literature review and meta-analysis, initial blood specimen diversion devices were associated with a lower frequency of blood culture contamination.Adopting diversion devices may reduce unwanted outcomes, such as unnecessary antibiotic use, prolonged hospital LOS, and mortality.Diversion devices also reduce costs associated with contamination and diagnostic error in healthcare settings.
Diagnostic stewardship involves the responsible use of diagnostic tests to improve patient outcomes while minimizing potential harms to patients and reducing healthcare cost.Diversion devices represent a useful diagnostic stewardship strategy [15,32].However, the implementation of diversion devices was not tested in scenarios without standard measures to reduce blood culture contamination (skin and bottle preparation, sterile collection technique, education, and training).Diversion devices should be implemented in conjunction with these strategies to reduce blood culture contamination.
To minimize contamination in blood cultures, a range of methods can be utilized, including proper training of personnel, skin disinfection before venipuncture, avoiding drawing blood from catheters unless necessary, using aseptic techniques, and utilizing prepackaged blood culture kits.According to the Clinical and Laboratory Standards Institute guidelines, the use of paired culture sets is recommended to distinguish between contaminant organisms and true pathogens.For the initial evaluation, it is suggested that four 10-mL bottles (equivalent to 2 sets) should be applied, in order to detect approximately 90%-95% of bacteremias, whereas the bacteremia detection rate of 3 sets is 95%-99% [33].It is also recommended to regularly monitor contamination rates and report surveillance data to phlebotomists and nurses.However, despite these measures, some microbes may still survive local skin disinfection and accidentally contaminate the blood sample.
Blood culture contamination has many negative impacts on patients.It can lead to misdiagnosis and subsequent inappropriate use of antibiotics, either by prescribing the wrong antibiotic or by extending duration, contributing to the development of antimicrobial resistance.All these events are associated with prolonged hospital stay, which can increase the risk of healthcare-associated infections and complications.Patients may be required to do additional tests, leading to increased healthcare costs and overloading the healthcare system.In this scenario, the adoption of diversion devices could be a valid instrument to reduce untoward outcomes for patients and the healthcare system.
An important finding, and also limitation, of our study is the absence of a consensual definition for blood culture contamination.Two studies by the same research group [30,31] adopted the same blood culture contamination definition, while another group of 3 studies [20,28,29] followed similar definitions for blood culture contamination among them.Clearly there are some tendencies in the inclusion of some species in the definition, such as CoNS, Micrococcus spp, Corynebacterium spp, α-hemolytic viridans group streptococci, and Bacillus spp.However, we believe that a standardization in the definition of blood culture contamination would help the analysis of future studies, making them more methodologically homogeneous to compare.Since the result of a false-positive culture necessarily involves the definition of contamination, the lack of standardization interfered in our comparative analysis and continues to obscure the real extent of blood culture contamination [34].The College of American Pathologists proposed a definition of blood culture contamination, which is the growth of 1 or more of the following organisms in only 1 blood culture set and only 1 of a series of blood culture sets (eg, 1 of 1, 1 of 2, 1 of 3 sets): CoNS, Micrococcus spp, viridans group streptococci, Propionibacterium acnes, Corynebacterium spp, and Bacillus spp [2,9,35].Of the studies that we analyzed, only one-quarter followed this definition [20,28,29].More recently, the Veterans Health Administration microbiology group, with the objective of standardizing the definition of blood culture contamination, published a list of microorganisms whose growth in the culture bottle would be a sign of false-positive result and not true infection [36].
There are only a few diversion devices available on the market.The studies we included analyzed Steripath and Kurin.Despite having similar conclusions of diversion device benefits, Steripath was analyzed by studies of higher quality and clearer methodology than Kurin.An evident advantage of Kurin is the lower diverted blood volume, only 0.15 mL, in comparison to 1.5-2 mL of Steripath.Only 1 of the 3 studies that evaluated Kurin could be included in our meta-analysis since the other 2 did not report the total number of blood cultures, which was essential for the type of analysis we proposed to do.Diversion devices were capable of reducing blood culture contamination and, at the same time, of identifying true infection.In other words, the diversion of the initial blood specimen avoids contamination but does not hinder the detection of a true infection.
Our study has several limitations, mostly due to inherent gaps or constraints in the available primary literature.First, most studies (10) were not randomized [20][21][22][23][24][26][27][28][29][30].However, this is the most common study design in the infection prevention literature [37].Second, most studies were of moderate to low quality and may have overestimated or underestimated the results of blood culture contamination and true infection.Third, there were missing data of important analyzed aspects, such as antisepsis procedure, blood culture contamination definition, and outcomes (antibiotic use, hospital LOS, mortality).Furthermore, the lack of a standard blood culture contamination definition may have affected the homogeneity of our analysis.The quality of the reports of the studies varied, with some studies providing incomplete information about their methods and results.This may have introduced reporting bias into our analysis.However, we present a stratified analysis including high-quality studies showing that diversion devices can prevent blood culture contamination.Fourth, the number of blood cultures was considerably different among the studies, varying from 671 [30] to 41 685 [21], resulting in different weights during some of the stratified analysis.Fifth, some studies inferred outcomes (antibiotic use, hospital LOS, mortality), but not many analyzed them based on their own participants' data, which precluded our initial intention of analyzing the cost-effectiveness of diversion device adoption.Sixth, the studies included in our analysis were conducted in various settings, including different hospitals, and patient populations, which may have introduced heterogeneity and reduced the generalizability of our findings.Seventh, we only included studies that used a diversion device as the intervention and did not compare diversion devices to other methods of preventing blood culture contamination.This may have limited our ability to draw conclusions about the effectiveness of diversion devices relative to other interventions.Eighth, it was not part of the scope of this systematic literature review to study potential barriers to the adoption of diversion devices, such as staff resistance to change, the need for additional training, and whether the cost of the device systematically adopted is lower than the costs involved in the management of false-positive cases.It also should be noted that despite the limitations of the included studies, the use of a diversion device is not likely to mislead the true diagnosis in blood cultures.However, the studies may not have been designed to specifically address this question, and more research may need to confirm this finding.
Further research is needed to address several important unanswered questions.Specifically, head-to-head studies comparing different diversion devices would provide valuable insights into their relative effectiveness.Studies involving children would help clarify the possibilities and difficulties of applying diversion devices in that population.Additionally, conducting more robust cost-effectiveness studies would help determine the economic impact of implementing diversion devices compared to managing false-positive cases.
In conclusion, the use of diversion devices in healthcare facilities could be a valid instrument to reduce unnecessary antibiotic use and unwanted outcomes to patients and to the healthcare system.

Figure 1 .
Figure 1.Literature search for articles that evaluated the impact of a blood diversion device in culture contamination.Abbreviation: Cumulative Index to Nursing and Allied Health Literature.

Figure 2 .
Figure 2. Forest plot of blood culture contamination with a diversion device or a standard procedure of blood collection, in high-quality (Downs and Black ≥18) studies.Odds ratios were determined with the Mantel-Haenszel random-effects method.Abbreviations: CI, confidence interval; M-H, Mantel-Haenszel.

Figure 3 .
Figure 3. Forest plot of true infection detection with a diversion device or a standard procedure of blood collection.Odds ratios were determined with the Mantel-Haenszel random-effects method.Abbreviations: CI, confidence interval; M-H, Mantel-Haenszel.