Turnaround times of pathology results (from collection through to clinical interaction/issuing a report) have a significant impact on individual patient management, but also have a wider bearing on infection control/public health, hospital patient flows and antibiotic stewardship. Although blood cultures are collected from among the sickest patients, they are rarely treated as urgent. Without audit of the blood culture pathway (using specimen collection as the starting point), microbiologists and clinicians are unaware of significant preventable delays in obtaining results.
Over 30 years ago, Holliman et al. highlighted the need for rapid microbiology results, reporting that antibiotic treatment was either initiated or altered on the basis of laboratory results in half of patients with significant positive cultures . This was in an era where resistance to third-generation cephalo-sporins, quinolones and aminoglycosides was uncommon. Since then, antibiotic resistance rates have increased in clinical isolates, culminating in the emergence of carbapenamase-resistant Enterobacteriaceae. Blood culture technology has improved over the past decades, with laboratories now using analysers that monitor samples every 10e15 min and detect positive cultures 24 h/day. However, these developments have not been matched by changes in laboratory practice.
The present-day convergence of the need for improved recognition and management of sepsis, increasing antibiotic resistance, and the need for enhanced antibiotic stewardship places greater demands on the laboratory for improved turn-around times of blood cultures, with both positive and negative cultures having an impact on patient management.
The authors devised an optimized blood culture pathway in the study hospital. This study investigated the impact of this pathway on the turnaround times of results, and compared the blood culture turnaround times at the study hospital with those of five other laboratories that had not optimized their path-ways, and one laboratory that had taken some steps to improve blood culture handling and processing.
Optimization of the blood culture pathway
Addressing load delays
The guidelines of the UK Standards for Microbiology In-vestigations indicate that 100% of blood cultures should be loaded within 4 h of collection . A baseline audit prior to the intervention revealed that more than 60% of blood cultures were taking >4 h to be loaded at the study hospital. This was corrected in three stages:
Moving the FX blood culture analyser (Becton Dickinson, Oxford, UK) from microbiology into the blood sciences laboratory, allowed blood cultures to be loaded 24 h/day.
Replacing glass bottles with plastic blood culture bottles, allowing samples to be sent via the hospital air tube system.
Education of clinical staff on the importance of collecting and sending blood cultures to the laboratory without delay.
Addressing unload delays
Blood sciences staff processed blood cultures, flagging positive cultures outside of routine microbiology hours (08:30e20:00 h). Samples were plated on to routine laboratory media, including plates for direct Gram-negative sensitivity testing, extended-spectrum beta-lactamase testing and genta-micin minimum inhibitory concentration determination, in a portable class I safety cabinet. A Gram stain was not performed.
Audit of blood culture processing in other centres
Audit of other hospitals
Laboratories serving five other hospitals (teaching and non-teaching, some off-site) in the same health region as the study hospital provided the following data points on 27 consecutive Escherichia coli-positive blood cultures:
time when blood culture collected;
time when loaded on the analyser;
time when flagged positive; and
time when removed from the analyser.
None of these laboratories had optimized their blood cul-ture pathways. In addition, the same data set was collected for 50 consecutive blood cultures positive for E. coli at one other hospital (with an on-site laboratory) in another health region; a further data point (time when sensitivity data were inputted into the laboratory information management system) was also measured. These data were compared with the same time points in the study hospital.
Appropriateness of empirical antibiotic therapy
Using a pro forma, the initial antibiotic therapy of 106 consecutive patients with significant positive blood cultures was reviewed. Antibiotic therapy was considered to be appropriate if the patient was prescribed at least one agent that was active against the blood culture isolate based on in-vitro antibiotic susceptibility testing. For deep-seated in-fections due to Staphylococcus aureus, agents with modest activity (e.g. co-amoxiclav) were considered as partial ther-apy. Inappropriate therapy was defined as no antibiotic treat-ment, any oral antibiotic therapy in a patient who was septic, or parenteral treatment with antibiotics to which the pathogen was resistant. The authors also considered whether Gram stain results, or identity of the organism, before antibiotic suscep-tibilities were available could have corrected inappropriate empirical therapy.
Blood cultures positive for E. coli
Ninety-five percent of blood cultures were loaded within 2 h at the study hospital. In contrast, in the non-optimized hospi-tals, blood culture samples sometimes took over 24 h to load after collection 95% of bottles, with a range of 16e26 h (Figure 1). Ninety-seven percent of cultures positive for E. coli at the study hospital were removed from the analyser within 18 h of collection, compared with 42e56 h in the other hospitals (Figure 1). The average time from collection to unloading at the study hospital was 12.79 h, compared with 18.87e30.28 h in the other hospitals.
The study hospital was also substantially quicker than the comparator hospital that had optimized its blood culture pathway at three defined time points. The overall impact of this was that >85% of blood cultures positive for E. coli had antibiotic susceptibilities reported within 36 h of specimen collection at the study hospital, compared with 66 h at the comparator hospital (Figure 2).
Appropriateness of empirical antibiotic therapy
Of 106 consecutive significant positive blood cultures, almost one-third (N¼34) of patients did not receive appro-priate empiric antibiotic therapy. Analysis of failure of initial empirical therapy showed that a Gram stain result could have corrected treatment in 19 (55.9%) cases, and enabled early identification of the organism in a further five (14.7%) cases. Early availability of antibiotic susceptibilities would have influenced treatment in 10 (29.4%) cases.
Optimization of the blood culture pathway reduced turn-around times substantially. By optimizing the pathway, the authors were able to report 36-h negative blood cultures for neonates, in accordance with National Institute for Health and
Care Excellence guidance . This contrasts with the experi-ence in one UK region, where no laboratory could consistently provide real-time 36-h negative reports .
Reduction in load delays was achieved with minimal in-vestment because, as in most hospitals, blood sciences staff at the study hospital operate a 24-h shift system. Also, although
the blood science staff at the study hospital entered complete patient demographics into the microbiology laboratory infor-mation management system, modern analysers allow anony-mous loading, so this step could be eliminated.
Unload delays can be at least as long as load delays if blood cultures flagging positive outside routine hours remain on the analyser until the next working day. A national survey in 2012 found that some laboratories stopped processing positive blood cultures after 16:00 h during the working week and 12.00 at weekends . Five years later, these practices continue in some UK laboratories. One potential impact of delays at this stage is that some bacteria, most notoriously Streptococcus pneumoniae, may complete their growth cycle and autolyse so that subcultures are negative. However, the main benefit of early results availability is facilitating appropriate antibiotic therapy. This study found that empiric antibiotic treatment was inappropriate in almost one-third of cases of true blood-stream infection, although, perhaps surprisingly, early anti-biotic susceptibility data influenced treatment in only 10 out of 106 cases. However, this study did not evaluate the impact of earlier blood culture results on rationalizing or stopping antibiotic therapy.
There is a belief that all patients with positive blood cul-tures are on appropriate antibiotics, thereby mitigating the impact of delays. Such confidence is misplaced, as intrinsic to any antibiotic policy is the inevitability that it will not provide 100% activity against Gram-negative pathogens. Other studies, including a recent UK multi-centre study of blood-culture-confirmed Gram-negative sepsis, have found that approxi-mately one-third of patients do not receive appropriate empiric therapy [5,6] The present study, based in a district general hospital, may understate the size of the problem in other hospitals where patients may have infection with a wider range of micro-organisms, and where antibiotic susceptibilities may be more unpredictable. In the study hospital, making an incorrect initial diagnosis or failing to follow antibiotic pre-scribing guidelines were more common reasons for the inap-propriateness of empiric antibiotic therapy than unexpected antibiotic resistance.