Contamination Frustration: Considering Value-Based Approaches to Manage Contaminated Cultures

A 7-week-old, healthy, ex–full-term infant presented to the emergency department (ED) with fever in the setting of cough and rhinorrhea. Her review of systems was otherwise negative, and her physical exam was notable only for rhinorrhea without focal findings of bacterial infection. Laboratories performed per Pediatric Emergency Care Applied Research Network recommendations revealed no leukocytosis, with a normal absolute neutrophil count, procalcitonin of 0.12 ng/mL, and an unremarkable urinalysis. Blood cultures were sent, lumbar puncture deferred, and she was discharged from the hospital without empirical antibiotics. At 38 hours of incubation, the blood cultures grew gram-positive cocci in clusters. The ED provider instructed the family to bring the patient in for evaluation. Interim history revealed improving cough and congestion, and resolution of fever. Reevaluation in the ED confirmed a well-appearing, afebrile infant without signs of bacterial infection. Repeat blood cultures were sent and the patient was admitted for observation without initiation of antibiotics.

The inpatient team admitted the patient for a planned 24 hour monitoring period, continuing observation even after the initial blood culture speciated as coagulase-negative Staphylococcus (CoNS). The patient remained afebrile and well-appearing, but as the team was planning hospital discharge, the laboratory called to report that the repeat culture was now growing gram-positive cocci in clusters at ∼ 28 hours of incubation. Inpatient observation of the patient continued without antibiotics until the repeat blood culture also speciated as CoNS at 43 hours. This repeat culture was deemed another contaminant and the patient was eventually discharged from the hospital with no further workup or intervention. On pediatrician follow-up, there were no reported sequelae of bacterial infection.

This case represents the repercussions of false-positive results and offers an opportunity to consider alternative approaches that may have avoided potential patient harm and excessive health care costs. In the setting of a positive blood culture, regardless of clinical assessment, most clinicians feel obligated to repeat the culture. This case calls that general approach into question. The patient in this case was classified as low risk on the basis of Pediatric Emergency Care Applied Research Network, which placed her at a <0.2% chance of invasive bacterial infection (IBI). She never received antibiotics at her initial evaluation, eliminating the potential for a partially treated IBI. Finally, she was clinically well without a fever at the time of reevaluation in the ED, further reducing the already miniscule pretest probability for IBI, specifically Staph aureus in this case of GPC in clusters, to a level well below most clinicians’ threshold for diagnostic evaluation. If they hadn’t known of the earlier positive culture, the providers likely would not pursue a blood culture in an afebrile child with a resolving viral infection. Should the positive culture have affected clinical decision-making?

Previous work has investigated factors associated with increased likelihood of a contaminated blood culture, though much of this work excludes the infant population. The likelihood of contaminant is higher when multiple culture sets are obtained and only a portion of them grow bacteria.⁶  However, almost 89% to 100% of infants have only 1 set of cultures sent, likely because of patient discomfort and blood volume required to obtain multiple sets.⁷  When only 1 set is sent, providers often consider whether 1 or both bottles grew bacteria as another marker for likelihood of contamination, but this has been shown to be an unreliable marker for contamination given that many bacteria simply grow better in either aerobic or anaerobic environments.⁸  Another unreliable marker because of limited data in pediatrics is the quantity of growth per culture bottle because cultures with low colony-forming units per milliliter may not necessarily indicate contaminant over true pathogen.⁹ 

Given the limitations of these factors, time to positivity (TTP) remains 1 of the most useful tools in predicting contaminated culture results within pediatrics. In 256 infants <60 days old, the median TTP for true pathogens was 16.6 hours versus 25.1 hours for contaminants.¹⁰  In infants aged 0 to 90 days, the likelihood of contaminant was 7.8 times higher for each culture that resulted positive after 36 hours.¹¹  Similarly, in febrile children aged 3 months to 3 years, true pathogens were shown to grow quicker at 13.8 ± 7.0 hours versus 37.6 ± 29.9 hours for contaminated cultures.¹²  In another study of 10 200 pediatric blood cultures, the TTP of true pathogens was 18.4 hours compared with 32.8 hours for contaminants.¹³  When looking specifically at CoNS-positive blood cultures in pediatric patients, a TTP of ≤15 hours has been shown to have a positive predictive value of 84% for pathogenicity, whereas a TTP of ≥22 hours had a positive predictive value of 87% for contaminant culture.¹⁴  For the case presented here, the delayed TTP coupled with the low likelihood of IBI suggested the result was highly likely to be a contaminant.

Repeating the blood culture led to many downstream effects, including multiple days of inpatient hospitalization, repeat laboratory testing, and likely significant psychosocial stress on the parents of the newborn. Contaminated pediatric blood cultures have been shown to significantly increase length of stay and overall hospital charges.^8,15  When antibiotics are initiated, patients are submitted to unnecessary courses of antibiotics and other iatrogenic exposures, including intravenous infiltrations and allergic reactions. The financial burden is also significant, including the cost of prolonged inpatient stays, laboratory charges, and, if antibiotics are started, pharmacy and medication costs. Finally, hospitalizations in young infants have also been associated with vulnerable child syndrome, which increases the patient’s risk for mood, sleep, and performance issues later in life^16,17 

As with all tests, clinicians must weigh the pretest probability of a disease to determine if the risks of testing are in the patient’s best interest. In our patient’s case, the increased likelihood of a contaminant, coupled with the low pretest probability for IBI, seem to tip the scales toward risk over benefit. A conservative approach might have included admission of the patient without repeating the blood culture while awaiting speciation, which would have shortened the initial hospitalization and avoided the second false-positive culture result. This option is even more appealing when considering advancements in polymerase chain reaction tests that can help rule out meticillin-sensitive Staphylococcus aureus/meticillin-resistant Staphylococcus aureus in cases of GPC such as this one.⁸  The clinicians could have also reasonably opted to send the patient home with follow-up, with counseling to monitor for any concerning changes in a recovering well-appearing infant.

Once admitted, the inpatient team could have discharged her from the hospital with knowledge of CoNS from the initial culture and not monitored the other culture. Once the second culture turned positive, the team could have applied the logic described above and had the patient monitored in the outpatient setting. Though the presence of the same pathogen in the repeat culture might raise suspicion for a true infection, the persistent well-appearance offers significant reassurance against that possibility.

In addition to standardizing practices around contaminant cultures, it is crucial to continue to reduce contamination rates. Many studies have shown that the implementation of quality improvement interventions such as nursing checklists and staff education, phlebotomy teams and blood culture collection kits, and optimization of blood culture volume and specimen handling decreased contamination rates and cost burden.^3,18–20  In addition to implementing interventions from a collection and handling standpoint, efforts to reduce rates of false-positive cultures rely on sending cultures only in appropriate settings.²¹  Though the first culture in our case certainly represents an appropriate setting, the repeat culture could likely have been avoided.

Contaminated culture results continue to be a source of clinical frustration and there are no formal guidelines/evidence-based recommendations to help guide clinical approaches, which leads to variation in care. These decisions become even more difficult when TTP is <36 hours, patients have initially been exposed to antibiotics, or when multiple sets of cultures have positive results for classically contaminant species. In lieu of clear guidance, clinicians must weigh the likelihood of true disease with the risks of repeating a blood culture and avoid the reflex to simply repeat cultures, provide empirical antibiotics, and admit the patient to the hospital. This is where providers can take control of minimizing interventions and potential downstream harm to the patient.

We thank the patient and patient’s family described in this case for allowing us to present their story.