Obtaining Blood Cultures by Venipuncture versus from Central Lines: Impact on Blood Culture Contamination Rates and Potential Effect on Central Line–Associated Bloodstream Infection Reporting
Objective. Reduce the frequency of contaminated blood cultures that meet National Healthcare Safety Network definitions for a central line–associated bloodstream infection (CLABSI).
Design. An observational study.
Setting. A 500-bed university-affiliated hospital.
Methods. A new blood culture policy discouraged drawing blood samples from central lines. Phlebotomists were reeducated regarding aseptic technique when obtaining blood samples by venipuncture. The intravenous therapy team was taught how to draw blood samples by venipuncture and served as a backup when phlebotomists were unable to obtain blood samples. A 2-nurse protocol and a special supply kit for obtaining blood samples from catheters were developed. Rates of blood culture contamination were monitored by the microbiology laboratory.
Results. The proportion of blood samples obtained for culture from central lines decreased from 10.9% during January–June 2010 to 0.4% during July–December 2012 (). The proportion of blood cultures that were contaminated decreased from 84 (1.6%) of 5,274 during January–June 2010 to 21 (0.5%) of 4,245 during January–June 2012 (). Based on estimated excess hospital costs of $3,000 per contaminated blood culture, the reduction in blood culture contaminants yielded an estimated annualized savings of $378,000 in 2012 when compared to 2010. In mid-2010, 3 (30%) of 10 reported CLABSIs were suspected to represent blood culture contamination compared with none of 6 CLABSIs reported from mid-November 2010 through June 2012 ().
Conclusions. Multiple interventions resulted in a reduction in blood culture contamination rates and substantial cost savings to the hospital, and they may have reduced the number of reportable CLABSIs.
Central line–associated bloodstream infections (CLABSIs), once considered to be an unavoidable consequence of complex medical care, have been shown to be largely preventable.1 Following implementation of mandatory reporting of CLABSIs, considerable debate has arisen regarding the accuracy of the surveillance definitions used for reporting CLABSIs.2-5 For example, infectious diseases specialists at one hospital classified 45% of 20 reported CLABSIs as most likely representing contaminated blood cultures rather than true CLABSIs.3 The authors concluded that better aseptic technique for obtaining blood samples for culture could lower the number of reportable CLABSI cases. In late 2010, a review by the hospital epidemiologist of 10 recently reported CLABSIs at the Hospital of Saint Raphael (now Saint Raphael campus of Yale-New Haven Hospital) suggested that positive blood culture results from 3 cases (30%) most likely represented contaminated cultures of blood drawn from central catheters, even though these cases met National Healthcare Safety Network (NHSN) surveillance criteria for a CLABSI. As a result, the hospital implemented strategies to minimize the number of blood samples drawn from catheters. This report describes the impact of reducing the use of catheter-drawn blood samples for culture on blood culture contamination rates and its possible contribution to reducing the number of reportable CLABSIs.
In 2009, the hospital’s multidisciplinary committee for reducing CLABSIs focused its efforts in medical intensive care units (MICUs). Because CLABSIs continued to occur at an unacceptable rate, in late 2010, the hospital restructured and expanded its CLABSI reduction program.6 Based on Centers for Disease Control and Prevention and Johns Hopkins Hospital recommendations,7,8 the committee developed new policies that recommended drawing blood samples for culture by venipuncture whenever possible and avoiding the use of catheter-drawn blood samples for culture unless absolutely necessary. The hospital’s phlebotomy team, which obtains specimens for blood cultures only from peripheral veins, was reeducated about the importance of applying the alcoholic chlorhexidine antiseptic for the appropriate time and allowing the area to air dry. Because phlebotomists were sometimes unable to obtain peripheral venous blood samples for culture, the hospital’s intravenous (IV) team was also trained on how to draw blood samples by venipuncture and were asked to do so before requesting that intensive care unit (ICU) nurses obtain blood specimens via central catheters. Nursing personnel in ICUs where the CLABSI reduction program was initiated developed a written procedure designed to minimize contamination of blood specimens drawn from central catheters when neither phlebotomists nor the IV team could obtain blood specimens by venipuncture. The procedure required 2 nurses; one nurse obtained blood specimens, while the other monitored the procedure using a checklist. The procedure included removing needleless connectors and “scrubbing the hub” before obtaining specimens for culture.9,10 Nurses also developed a special kit that contains all necessary items for drawing blood specimens from catheters using the 2-nurse procedure (Table 1).
|Contents of kit|
|1 sterile package wrapping|
|1 syringe with cannula|
|1 end cap|
|1 needleless connector|
|6 alcohol prep pads|
|1 blood transfer device|
|1 specimen bag|
|2 prefilled sodium chloride syringes|
In mid-November 2010, a memorandum sent to infectious diseases attending physicians and MICU attending physicians recommended that blood samples for culture be obtained by venipuncture whenever possible. In mid-January 2011, a similar memorandum was sent by the hospital epidemiologist, the Director of the MICU, and the Chief Medical Officer to all members of the medical staff. Physicians were required to obtain permission from the hospital epidemiologist to have blood samples drawn for culture from central catheters unless the patient was febrile and neutropenic or required hemodialysis. In December 2011, the multidisciplinary, multimodal program to reduce CLABSIs was introduced hospital-wide.
At the time blood samples were obtained for culture, physicians or physician assistants were prompted to enter into the hospital’s clinical information system whether the blood was drawn from a central line or from other sites (peripheral vein or A-line). If a blood sample could not be obtained by venipuncture, then the protocol required that the order be cancelled and a new order placed for blood culture samples to be drawn from a catheter. The clinical microbiology laboratory developed a monthly report of the number of blood culture samples drawn on all units of the hospital. The report also included the proportion of blood cultures with samples drawn from central lines versus from other sites. The report was presented to the CLABSI committee on a regular basis. The number of blood cultures obtained each month was expressed as the number of blood cultures per 1,000 patient-days.
Positive blood cultures from patients with central lines were defined as indicating CLABSI by the infection control program using NHSN criteria.7 Only CLABSIs occurring in ICU patients were required to be reported as CLABSIs to NHSN. Blood culture contamination rates were determined after review of blood cultures by the same clinical microbiologist using published criteria for the entire study period.11 Microorganisms most likely to represent contaminants included coagulase-negative staphylococci, Corynebacterium species, Bacillus species, Propionobacterium acnes, Micrococcus, viridians streptococci, and enterococci.11,12 Other criteria used in classifying isolates as likely contaminants included the number of blood culture sets that were positive, the number of bottles that were positive per set, and the time to positivity.13 Blood culture contamination rates were determined for cultures obtained on all hospital units except for the emergency department and the neonatal intensive care unit (NICU), because phlebotomists do not draw blood samples for these 2 units. To estimate potential cost savings achieved by having fewer contaminated blood cultures, we used an excess cost of $3,000 per contaminated blood culture, which is an average of published cost estimates, which have ranged from $1,000 to $3,750 per contaminated blood culture.14-18
Differences in the proportions were compared by using χ2 or χ2 for trend analysis (EpiInfo, ver. 7). The number of blood cultures per 1,000 patient-days obtained throughout the hospital and for ICUs only, by 6-month period, was compared by use of the Z test (MedCalc).
The total number of blood cultures obtained in all patient care areas and the proportion obtained from central lines by quarter, from January 2010 through December 2012, are shown in Figure 1. The proportion of blood cultures obtained from central lines averaged 10.9% (range, 10.6%–11.2%) during the period January–June 2010 and was similar (mean, 10.5%; range, 9.9%–10.9%) during July–September 2010. As a result of development and initial dissemination of new policies recommending avoidance of drawing blood samples from central catheters in late October and early November, the proportion of blood culture samples obtained from catheters decreased significantly to an average of 6.4% in the last quarter of 2010 (; Figure 1). Following dissemination of the new blood culture policies to all medical staff in mid-January 2011, the proportion of blood culture samples drawn from catheters continued to decrease progressively from an average of 2.6% in the first 6 months of 2011 to an average of 0.4% in the last 6 months of 2012 (). Analysis of data limited to the ICUs revealed that the proportion of blood cultures obtained from central lines decreased from 720 (28.9%) of 2,488 in 2010 to 60 (2.5%) of 2,375 in 2011 and 16 (0.8%) of 2,035 in 2012 ().
The hospital-wide blood culture specimen collection rates increased gradually from 139 blood cultures per 1,000 patient-days during January–June 2010 to 157 blood cultures per 1,000 patient-days in the period July–December 2012 (). Analysis of data only from the ICUs revealed that the blood culture specimen collection rates decreased from 326 blood cultures per 1,000 patient-days in January–June 2010 to 287 blood cultures per 1,000 patient-days in July–December 2012 ().
Trends in the frequency of blood culture contamination for specimens obtained from patients on all nursing units except the emergency department and the NICU are shown in Table 2. The proportion of catheter-drawn blood cultures that were contaminated decreased progressively from 44 (4.8%) of 921 in the first 6 months of 2010 to 0 (0%) of 39 in the first 6 months of 2012 (; Table 2). The proportion of blood culture specimens obtained by venipuncture that were contaminated decreased from 40 (0.9%) of 4,353 during January–June 2010 to 21 (0.5%) of 4,206 during January–June 2012 (). The proportion of all contaminated blood culture specimens that were drawn from catheters decreased from 44 (52.4%) of 84 in the first 6 months of 2010 to 3 (7.1%) of 42 in early 2011 and 0 (0%) of 21 in the first 6 months of 2012 (; Table 2). Overall, the proportion of all blood culture specimens (drawn from either catheters or by venipuncture) that were contaminated decreased from 84 (1.6%) of 5,274 in the first 6 months of 2010 to 21 (0.5%) of 4,245 in the first 6 months of 2012 ().
|No. of contaminated blood culture specimens obtained by venipuncture||40||39||21|
|Total blood culture specimens obtained by venipuncture||4,353||4,733||4,206|
|Contamination of blood culture specimens obtained by venipuncture, %||0.9||0.8||0.5|
|No. of contaminated blood cultures obtained from catheters||44||3||0|
|Total blood culture specimens obtained from catheters||921||227||39|
|Contamination of blood culture specimens obtained from catheters, %||4.8||1.3||0|
|Total no. of contaminated blood culture specimens||84||42||21|
|Total no. of blood culture specimens obtained||5,274||4,960||4,245|
|Contamination of all blood culture specimens, %||1.6||0.8||0.5|
When analysis of blood culture contamination rates was limited to the ICUs, we found that the proportion of catheter-drawn blood culture specimens that were contaminated decreased from 21 (5%) of 422 in the first 6 months of 2010 to 0 of 44 in the first 6 months of 2011 and to 0 of 9 in the first 6 months of 2012 (). The proportion of blood cultures obtained in ICUs by venipuncture that were contaminated remained relatively constant: 7 (0.8%) of 869 during January–June 2010, 5 (0.4%) of 1,161 during January–June 2011, and 7 (0.7%) of 1,019 during January–June 2012 (). The proportion of all contaminated blood culture specimens in ICUs that were drawn from catheters decreased from 21 (75%) of 28 in the first 6 months of 2010, to 0 of 5 in early 2011, and to 0 of 7 in the first 6 months of 2012 ().
Based on estimated excess hospital costs of $3,000 per contaminated blood culture, the 84 contaminated blood cultures that occurred in the first 6 months of 2010 were estimated to have generated excess costs of $252,000 for the 6-month period, yielding an estimated annualized excess cost of $504,000. In contrast, the 21 contaminated blood cultures that occurred in the first 6 months of 2012 were estimated to have generated excess costs of $63,000 and an estimated annualized cost in 2012 of $126,000. Compared with 2010, the estimated annual cost savings in 2012 from fewer blood culture contaminants was $378,000. If a very conservative excess cost of $1,000 per contaminated blood culture is used, the estimated annual savings in 2012, when compared with 2010, would be $126,000.
In mid-2010, 3 (30%) of 10 recently reported CLASBIs were suspected to represent blood culture contamination rather than true CLABSIs. One patient had a blood culture specimen drawn from a peripherally inserted central catheter (PICC) which grew Enterococcus faecalis in only 1 bottle, whereas 2 blood culture specimens drawn from hemodialysis catheters on the same day had negative results. Also, blood culture specimens drawn from the PICC 2 days earlier and 1 and 3 days later were negative. A second patient’s blood culture sample drawn from a triple-lumen catheter (TLC) grew Klebsiella pneumoniae in a single bottle, whereas a blood culture specimen obtained via venipuncture the same day was negative. A blood culture specimen drawn from the TLC the next day had negative results. Two blood culture specimens drawn from the TLC, 2 drawn from a hemodialysis catheter, and a semiquantitative culture of the TLC catheter tip the following day all had negative results. The patient received a diagnosis of Clostridium difficile infection and had clinically improved before treatment for Klebsiella infection was initiated. A third patient had a blood culture specimen drawn from a hemodialysis catheter that yielded Staphylococcus epidermidis in a single bottle, with a time to positivity of 26 hours. A blood culture specimen obtained via venipuncture the same day had negative results. The following day, the patient had another blood culture specimen drawn from the hemodialysis catheter that yielded S. epidermidis in a single bottle, with a time to positivity of 30.5 hours. A semiquantitative culture of the hemodialysis catheter the following day yielded no growth. In contrast, none of 6 CLABSIs reported during mid-November through June 2012 yielded findings suspected to represent contaminated blood cultures ().
As a result of numerous changes in hospital policies and procedures implemented by the multidisciplinary CLABSI reduction committee, the number of intensive care unit CLABSIs reported to the NHSN decreased from 17 in 2010 to 3 in 2012. The CLABSI rate decreased from 2.2 cases per 1,000 catheter-days in 2010 to 0.42 cases per 1,000 catheter-days in 2012 ().
We demonstrated that a combination of measures resulted in a progressive and sustained reduction in our blood culture contamination rate from 1.6% to 0.5% for all hospital units excluding the emergency department and NICU (). Implementing an institutional policy to minimize blood culture specimens drawn from central lines and training IV nurses to obtain blood culture specimens when the phlebotomy team was unsuccessful significantly reduced the proportion blood culture specimens that were drawn from central lines. Requiring permission from the hospital epidemiologist to draw blood culture specimens from catheters likely served as a significant barrier to physicians ordering cultures of blood specimens drawn from catheters. Limiting the number of blood culture specimens obtained from central lines undoubtedly contributed to reducing the blood culture contamination rate, which is not surprising, because it is well-established that cultures obtained from catheters have a higher contamination rate than do cultures obtained via venipuncture.19 In fact, blood culture specimens drawn from central lines accounted for a slight majority (44 [52.4%] of 84) of all contaminated blood cultures in early 2010 but accounted for none of the 21 contaminated blood cultures that occurred during January through June 2012.
The low blood culture contamination rate (1.6%) at the beginning of the observation period was similar to rates ranging from 1.1% to 3.1% noted in other facilities with phlebotomy teams.19 A recent meta-analysis and additional studies have found that blood culture contamination rates are significantly lower when cultures are routinely drawn by phlebotomists, as opposed to nonlaboratory personnel such as nurses, nursing aides, medical students, or physicians in training.12,19-22 For this reason, recent national guidelines recommend that, where available, a phlebotomy team should draw blood samples for culture.23 Reeducation of personnel responsible for drawing peripheral vein blood culture samples (in our study, phlebotomists) regarding the importance of aseptic technique24 and having a trained IV team draw blood cultures from patients with limited venous access likely contributed to the 44% reduction in the frequency with which blood cultures obtained from sites other than central lines (the vast majority were via venipuncture) were contaminated. Reducing the proportion of blood cultures drawn from central lines that were contaminated from 4.8% to less than 1% was likely attributable to the use of a written 2-nurse protocol and special kit when drawing blood culture samples from central lines. Although compliance with the 2-nurse protocol was not formally monitored, compliance rates as high as 100% were estimated. Relatively few blood culture samples were obtained from catheters after implementation of the protocol. Excluding blood culture samples obtained in the emergency department may have contributed to low contamination rates noted throughout the study period, because contamination rates have been shown to be higher in emergency departments.25 Some physicians initially expressed concern that obtaining blood culture specimens from catheters only in unusual situations in which they could not be obtained via venipuncture might result in patients not having cultures done when clinically indicated. We found that the hospital-wide rate of blood culture sample collection increased during the study period, which suggests that this was unlikely. Although the blood culture sample collection rate did decrease in the ICUs, we are unaware of any adverse clinical outcomes that could be attributed to the reduction in blood culture sample collections rates in the ICUs.
Multiple studies have found that contaminated blood cultures are associated with excess hospital costs and increased use of resources.14-18 Increased costs are attributable to greater use of laboratory tests, intravenous antibiotics, and microbiology charges.14 Some but not all studies have found that contaminated blood cultures may also result in an increased length of stay, further increasing hospital costs.14,16,18 The estimated annualized excess costs associated with contaminated blood culture samples in our study were lower than estimated costs or charges of more than $1 million per year reported by others.16-18
We could not determine the extent to which reducing blood culture contamination contributed to a reduction in the number of positive blood cultures that met NHSN criteria for a CLABSI, because multiple other changes relating to central line insertion, maintenance, and removal were made during the observation period.6 However, significantly reducing the number of contaminated blood culture specimens on all units except the emergency department and the NICU may have contributed in part to a reduction in reportable CLABSI cases, because positive blood cultures that appear, from a clinical perspective, to represent contaminants may meet NHSN criteria for CLABSI.3,10 We did observe that fewer CLABSIs reported in 2011 and 2012 than in 2010 had characteristics that were suggestive of blood culture contamination, although the difference did not reach statistical significance.
Our study has a number of limitations. It was an observational study conducted in a single 500-bed teaching hospital that cared for a limited number of high-risk patients, such as those with hematologic malignancies, bone marrow or solid-organ transplantation, and other conditions that require long-term central catheters. The level of cooperation of physicians in training and attending physicians regarding our policy to minimize drawing blood culture samples from central lines may have been greater than what might be achievable in much larger institutions. Requiring permission from a hospital epidemiologist or CLABSI committee chairman to order cultures of blood samples obtained from a central line may not be feasible in large institutions with many patients who have very limited peripheral venous access. Our ability to dramatically reduce the number of blood culture samples obtained from catheters and increase cultures obtained by venipuncture was facilitated by the availability of a phlebotomy team and IV team, which may not be present in other facilities. However, we do not feel that having a phlebotomy team and IV team is an extraordinary resource, based on recent guidelines and meta-analysis.19,23 The criteria used to classify positive blood cultures as contaminants may not be the same as in other institutions, but they were consistent throughout the study period. Our infection control program used NHSN criteria when determining which bloodstream infections were reportable as CLABSIs.7 However, a recent survey established that there is substantial variation between facilities in methods used to diagnose and report CLABSIs.5 Estimates of the cost savings achieved by having fewer contaminated blood cultures were based on published data, rather than on an analysis of direct and indirect costs associated with patients whose positive blood culture results were classified as due to contaminants. Also, we did not establish the impact of fewer contaminated blood cultures on the use of vancomycin, which is often given unnecessarily to patients with contaminated blood cultures.15
In conclusion, the interventions that we implemented achieved a significant and sustained reduction in the number of blood culture samples drawn from central lines and the number of contaminated blood cultures. The reduction in contaminated blood cultures likely resulted in cost savings for the hospital and may have had a beneficial impact on the number of CLABSIs reported.
Potential conflicts of interest. All authors report no conflicts of interest relevant to this article. All authors submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest, and the conflicts that the editors consider relevant to this article are disclosed here.
- 1. Pronovost P, Needham D, Berenholtz S, et al. An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med 2006;355:2725–2732.
- 2. Sexton DJ, Chen LF, Anderson DJ. Current definitions of central line–associated bloodstream infection: is the emperor wearing clothes? Infect Control Hosp Epidemiol 2010;31:1286–1289.
- 3. Shuman EK, Washer LL, Arndt JL, et al. Analysis of central line–associated bloodstream infections in the intensive care unit after implementation of central line bundles. Infect Control Hosp Epidemiol 2010;31:551–553.
- 4. Teplick R. Caveats for comparing catheter-associated bloodstream infection rates. Crit Care Med 2011;39:392–394.
- 5. Beekmann SE, Diekema DJ, Huskins WC, et al. Diagnosing and reporting of central line–associated bloodstream infections. Infect Control Hosp Epidemiol 2012;33:875–882.
- 6. Dumigan DG, Reilly L, Acompora M, et al. Beyond the “bundle”: interventions to decrease catheter-associated bloodstream infections in a community teaching hospital. Paper presented at: 22nd European Society of Clinical Microbiology and Infectious Diseases meeting; April 3, 2012; London.
- 9. Mathew A, Gaslin T, Dunning K, Ying J. Central catheter blood sampling: the impact of changing the needleless caps prior to collection. J Infus Nurs 2009;32:212–218.
- 10. Sherertz RJ, Karchmer TB, Palavecino E, Bischoff W. Blood drawn through valved catheter hub connectors carries a significant risk of contamination. Eur J Clin Microbiol Infect Dis 2011;30:1571–1577.
- 11. Hall KK, Lyman JA. Updated review of blood culture contamination. Clin Microbiol Rev 2006;19:788–802.
- 12. Weinstein MP. Blood culture contamination: persisting problems and partial progress. J Clin Microbiol 2003;41:2275–2278.
- 13. Dwivedi S, Bhalla R, Hoover DR, Weinstein MP. Discarding the initial aliquot of blood does not reduce contamination rates in intravenous-catheter-drawn blood cultures. J Clin Microbiol 2009;47:2950–2951.
- 14. Bates DW, Goldman L, Lee TH. Contaminant blood cultures and resource utilization. JAMA 1991;265:365–369.
- 15. Souvenir D, Anderson DE Jr, Palpant S, et al. Blood cultures positive for coagulase-negative staphylococci: antisepsis, pseudobacteremia, and therapy of patients. J Clin Microbiol 1998;36:1923–1926.
- 16. Zwang O, Albert RK. Analysis of strategies to improve cost effectiveness of blood cultures. J Hosp Med 2006;1:272–276.
- 17. Gander RM, Byrd L, DeCrescenzo M, Hirany S, Bowen M, Baughman J. Impact of blood cultures drawn by phlebotomy on contamination rates and health care costs in a hospital emergency department. J Clin Microbiol 2009;47:1021–1024.
- 18. Alahmadi YM, Aldeyab MA, McElnay JC, et al. Clinical and economic impact of contaminated blood cultures within the hospital setting. J Hosp Infect 2011;77:233–236.
- 19. Snyder SR, Favoretto AM, Baetz RA, et al. Effectiveness of practices to reduce blood culture contamination: a laboratory medicine best practices systematic review and meta-analysis. Clin Biochem 2012;45:999–1011.
- 20. Schifman RB, Strand CL, Meier FA, Howanitz PJ. Blood culture contamination: a College of American Pathologists Q-Probes study involving 640 institutions and 497,134 specimens from adult patients. Arch Pathol Lab Med 1998;122:216–221.
- 21. Weinbaum FI, Lavie S, Danek M, Sixsmith D, Heinrich GF, Mills SS. Doing it right the first time: quality improvement and the contaminant blood culture. J Clin Microbiol 1997;35:563–565.
- 22. Bekeris LG, Tworek JA, Walsh MK, Valenstein PN. Trends in blood culture contamination: a College of American Pathologists Q-Tracks study of 356 institutions. Arch Pathol Lab Med 2005;129:1222–1225.
- 23. Mermel LA, Allon M, Bouza E, et al. Clinical practice guidelines for the diagnosis and management of intravascular catheter-related infection: 2009 update by the Infectious Diseases Society of America. Clin Infect Dis 2009;49:1–45.
- 24. Eskira S, Gilad J, Schlaeffer P, et al. Reduction of blood culture contamination rate by an educational intervention. Clin Microbiol Infect 2006;12:818–821.
- 25. Suwanpimolkul G, Pongkumpai M, Suankratay C. A randomized trial of 2% chlorhexidine tincture compared with 10% aqueous povidone-iodine for venipuncture site disinfection: effects on blood culture contamination rates. J Infect 2008;56:354–359.