Best Way to Prevent Blood Culture Contamination

Becky A. Smith, MD; Lance R. Peterson, MD


November 17, 2014

Editorial Collaboration

Medscape &

Does Chlorhexidine Bathing in Adult Intensive Care Units Reduce Blood Culture Contamination? A Pragmatic Cluster-Randomized Trial

Septimus EJ, Hayden MK, Kleinman K, et al
Infect Control Hosp Epidemiol. 2014;35(Suppl 3):S17-S22

Strategies to Prevent Blood Culture Contamination

False-positive, or contaminated, blood cultures are recognized as a healthcare quality issue.[1] Their occurrence contributes to increased length of stay, medical cost, unnecessary antibiotics, and avoidable drug toxicity.[2] Target rates for contamination are 2%-3%,[2] with a median US adult inpatient rate of 2.5%.[3]

Septimus and colleagues conducted a prospective, pragmatic, cluster-randomized trial involving 74 intensive care units (ICUs) in 43 hospitals, 42 of which were community facilities. Three strategies of preventing blood culture contamination were compared:

Culture-based nasal surveillance for methicillin-resistant Staphylococcus aureus (MRSA), with isolation of positive patients (group 1);

Surveillance, with isolation and decolonization of positive patients (group 2); and

No surveillance, but universal decolonization with nasal mupirocin and daily chlorhexidine bathing (group 3).

During a 6-month baseline period, 7926 blood cultures were collected from 3399 patients: 32.3% in group 1, 27.3% in group 2, and 40.4% in group 3. During the 18-month intervention period, 22,761 cultures were collected from 9878 patients: 31% in group 1, 32.5% in group 2, and 36.5% in group 3.

Blood culture contamination was defined as having National Healthcare Safety Network (NHSN) commensal pathogens isolated from only one blood culture. Contamination rates for the three groups are shown in the Table.

Table. Blood Culture Contamination Rates, by Prevention Strategy

Period 1: Screening and Isolation 2: Targeted Decolonization 3: Universal Decolonization
Baseline 4.1% 3.9% 3.8%
Intervention 3.3% 3.2% 2.4%


Comparing full paired sets (eg, two samples) of cultures, group 3 had the greatest reduction in contamination rates (odds ratio, 0.55; 95% confidence interval [CI], 0.43-0.71).

Notable findings of this well-done study include the rigorous study design, which captured a broad swath of community hospital patients; results that highlight a significant secondary benefit from a new infection control initiative; and the challenge of contaminated blood cultures in the ICU setting, where samples are typically collected by nursing staff or physicians.


The main question that arises is how best to reduce blood culture contamination rates. Hall and Lyman[1] suggest 5 critical areas for reducing contamination:

Skin preparation;

Culture bottle preparation;

Percutaneous collection;

Phlebotomy teams; and

Commercial collection kit use.

The application of alcohol-impregnated catheter hub caps reduces catheter hub contamination,[4] which can be another source of false-positive cultures collected from venous catheters. The current study hospitals had implemented several measures directed at lowering contamination rates, including "scrubbing the hub."

At the four-hospital NorthShore University HealthSystem, we have similar culture collection policies. In addition, we use phlebotomy teams outside the ICUs and emergency departments (EDs) and provide real-time feedback to these latter areas whenever culture contamination occurs. During 2013, cultures by phlebotomy personnel (3277 sets) had a contamination rate of 0.46% (95% CI, 0.27%-0.78%) and 6427 ICU/ED sets had a 2.16% contamination rate (95% CI, 1.82%-2.55%).

The key lesson is that blood culture contamination can be improved. Every hospital can improve their performance.



Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: