Quality of blood culture testing - a survey in intensive care units and microbiological laboratories across four European countries

Blood culture (BC) testing before initiation of antimicrobial therapy is recommended as a standard of care in international sepsis guidelines [1] and has been shown to contribute to a decrease in ICU stay [2‐4]. Furthermore, BC testing is one of the cornerstones for antibiotic stewardship programs, which has been shown to reduce antibiotic overuse and costs in hospitalized patients [5, 6].

Beside limitations of BC testing, for example antibiotic/antimycotic treatment prior to sampling, low proportion of causative agents in the blood samples, and frequent fastidious or noncultivable organisms [7‐9], a high degree of standardization in microbiological laboratory (LAB) practice warrants for an overall positivity of approximately 30 to 40% in case of severe sepsis or septic shock [10]. In a recent large multicenter trial from Germany [11] 33% of patients with severe sepsis or septic shock had proven bacteremia. This is in contrast to a rate of only 9.6% of positive blood cultures observed in clinical practice in German ICUs aside from protocolized care [12] and underlines shortfalls in the preanalytic procedures in the ICU. Such shortfalls cover inadequate skin antisepsis and sampling techniques, as access via intravenous catheters, low blood volumes and low numbers of BC sets drawn for inoculation, prolonged time to incubation, suboptimal preincubation prior to automated cultivation at 37°C, which have a significant effect on the diagnostic yield [13‐15].

The numbers of BC sets processed per hospitalized patients are off particular importance. According to the case mix of the hospital, inoculation of 100 to 200 BC sets per 1,000 patient days is recommended [16, 17]. These numbers are, however, far from routine use, at least in Germany, where 55 BCs per 1,000 patient days were surveyed in 201 ICUs in 2009 in contrast to France, where 165 BCs per 1,000 patient days were quoted [18]. The 2010 annual report of the European Antimicrobial Resistance Surveillance Network (EARS-Net) specified only 12.1 BCs per 1,000 patient days in 37 hospitals in Germany, compared to 46.5 in 27 hospitals in France, 46.1 in 26 hospitals in the UK, and 70.7 BCs per 1,000 patient days in 22 hospitals in Italy [19]. In a recent study published by the National Reference Centre for Hospital Infections (NRZ), data of the German hospital nosocomial infection surveillance system (KISS) from 2006 were used to investigate the association between the frequency of blood cultures and central venous catheter-associated bloodstream infection (CVC-BSI) rates in 223 intensive care units (ICU) [20]. The median number of BC sets taken was 60 with a huge variation from 3.2 to 680 per 1,000 patient days. The authors concluded that if an external benchmarking of CVC-BSI rates is intended, an adjustment according to the BC frequency is necessary.

Reasons for the disregard of current guidelines have been identified, among others, in infrastructural aspects. The number of infections confirmed by LABs closely depends on the availability of closely located LABs [12], which sets a focus for future improvements of uniform customs and recommendations and of technical procedures on the preanalytic side of BC routine. Furthermore, there may be differences in the quality of BC testing between countries since the establishment of clinical microbiology and infectious disease departments vary substantially among European countries. Especially in Germany, patient-centered clinical microbiology is only a branch of laboratory medicine [21].

The aim of this qualitative survey was to assess the current practice in BC testing in ICUs and LABs across four European countries. Issues were technical aspects of the preanalytic course and an assessment of the current practice and their quality on the basis of individual perceptions among the staff and directors of ICUs versus LABs.

Some 138 interviews were conducted between September and November 2009 in 79 microbiological laboratories (LABs) and 59 intensive care units (ICUs) in France, Germany, Italy, and UK (Table 1). Pediatric and neonatal ICUs were excluded. Interviewees were ICU directors, ICU residents, ICU nurses, LAB directors, and LAB managers. The survey was carried out by an international agency (Advention BP, London, UK) on behalf of BD Diagnostics (Heidelberg, Germany). To uncover prevalent trends in thought and opinion, the interview panel was selected to fulfill a given quota, for example per country 10 to 20 ICUs and microbiological laboratories, respectively. Furthermore, the panel had to be balanced between BD Diagnostics (49.5%) and bioMérieux (Craponne, France) (50.5%) customers. Data were collected using semi-structured techniques for example individual in-depth personal telephone interviews. The interview guide included, among others, a list of general topics and open questions such as sepsis awareness and indication for BC testing, preanalytic procedures, sample transport and preincubation, and BC processing and communication of results (see Table S1 in Additional file 1). The response rate was 100 percent, since personal interviews have the potential to overcome the poor response rates of a questionnaire survey [22]. According to the requirements of the ethics committee of Jena University Hospital (Jena, Germany), the survey needed no ethical approval.

Blood culture testing is definitively the gold standard and primary test to evaluate patients with sepsis [24]. However, despite European efforts to standardize BC testing similar to the US Clinical and Laboratory Standards Institute (CLSI) guidelines [17], there are different perceptions regarding the performance of BC testing between the interviewees from four European countries in our survey.

The S2k guidelines of the German Sepsis Society (GSS) [25] (see Table S2 in Additional file 2), the Italian Progetto LaSER [26], and the Britain Saving Lives (NHS) guidelines [27] recommend ≥2 BC sets in case of sepsis suspicion, which is supported by the recent international guidelines of the Surviving Sepsis Campaign (SSC) [1], whereas the French National Society of Anaesthesia and Intensive Care (SFAR) give no recommendations.

Major challenges in BC testing are low rates of true positivity due to antibiotic pretreatment prior to blood withdrawal, suboptimal sample volume, an inadequate number of BC bottles cultured and delays in time to incubation.

In a French monocentric study, Vitrat-Hincky et al. found that only 45% of patients had adequate numbers of BC sets and only 13% had optimal sample volumes (that is ≥10 ml per bottle) [28]. The authors of a review on true-positive rate, contamination rate, and collected blood volume of BC bottles in five Belgian hospital laboratories found that more than one-third of the BC bottles handled were incorrectly filled, irrespective of the manufacturer of the blood culture vials [29].

In our survey, blood volumes collected per BC bottle varied considerably between countries with on average less than 10 ml per bottle (8.3 ml in France, Italy with 11.5 ml as an exception), though ICU staff is aware of the fact that BC positivity is proportional to the blood volume taken.

Differences in qualities of recommended blood sampling for BCs (number of sets and volume per bottle) may be partly explained by different responsibilities among the ICU staff. BC sampling is mainly carried out by physicians in the UK, by nurses in France and Italy, and by both in Germany.

In our survey, time to incubation of BCs ranged from 2 h in the UK and up to 20 h in German remote LABs. Limitations in transport times for BCs had been reported by Kerremans et al. in the Netherlands [30]. The median transport time in this study was 3.5 h, with 47% of cultures exceeding the recommended 4 h. Off-site location and type of clinical specialty were the most important predictors of long transport times. Cultures collected during weekend days or on wards at the largest distances from the laboratory were also associated with long transport times.

Considerable differences between countries were observed with regard to blood transport and storage prior to automated incubation in our survey. Delays in transport times were mainly due to different transport modes (that is, via van, porter, or pneumatic tube) and infrastructure. With Germany as an obvious exception, LABs are usually closely related to hospitals resulting in a transport time ≤4 h. Together with a general trend to store blood during closing times at room temperature, which accounts for a further delay of ≤12 h, up to 20 h time to incubation occurs in Germany. In consequence, up to 14% of German ICUs already have direct access to an on-site BC incubation device. The impact of immediate incubation of BCs delivered to the laboratory outside its hours of operation on turnaround times, antibiotic prescription practices, and patient outcomes was assessed by Kerremans et al. in a study from the Netherlands [31]. The authors found no difference in length of stay or hospital mortality, but immediate incubation of BCs outside laboratory hours reduced turnaround times and accelerated antibiotic switching.

Positive BC results are of paramount importance for patient management. Similarly to surgery, where the close cooperation with the pathologists of hospitals guarantees the intraoperative rapid section with immediate diagnosis within a few hours, BC results have to be considered as emergencies. It is therefore mandatory to notify a clinical professional (physician, nurse practitioner) responsible for the coordination of BC testing between LABs and ICUs.

Furthermore, since many patients are seen at an emergency department at first instance and initial BCs are taken there, it is the responsibility of the LABs to determine the location of the patient once the cultures are positive. Our survey shows that most LABs transmit preliminary results (that is, on Gram-staining behavior of the microorganisms grown in culture) via telephone, allowing clinical professionals to fine-tune the initial empiric antibiotic treatment. Final results, including ID/AST information, are mostly sent via facsimile or as a written letter report. This is due to the complex nature of the information (resistance-testing results for >10 antibiotics) and to time and cost reasons. Direct oral or face-to-face communication is established in all interviewees’ countries except Germany. However, improving communication of BC results (including negative results) have been shown to reduce antibiotic usage in neonatal intensive care units [32]. Telephone transmission of critical laboratory results may be inaccurate. However, a study by Howe et al. showed only minor transmission errors [33].

Our study has several limitations. First, aberrations from guidelines may notably in part be due to the general phenomenon that treatment recommendations in ICUs only poorly comply with practice recommendations: ICU directors perceive adherence to be higher than it actually is [34]. We did not perform an audit on order to assess actual practice. However, the results of this survey show that even perception of current BC practices in European ICUs is suboptimal. Second, the survey was qualitative in nature, so only semi-structured techniques with open questions were applied and respondents were not randomly selected and our findings are not representative. For instance, the proportion of BC sets processed in LABs is influenced by the case mix of ICUs. In addition, we have no quantitative data on preanalytic procedures, that is, contamination data, blood volume, and routine practice subsequent to inoculation of BC bottles. Furthermore, due to the exploratory outcome of our research, a statistical analysis was not performed and our data cannot be used to make generalizations. However, by providing insights into BC testing practices in European ICUs, our study generates ideas and hypotheses for later quantitative research. Finally, we did not assess knowledge and attitudes concerning interpretation of BC results and, more importantly, therapeutic consequences. However, guideline-based collection, processing and reporting of BCs are the cornerstones for successful patient management [35].

Evidence-based blood culture testing is of utmost importance for ICU patients with suspected sepsis. Knowledge of the etiologic agent (bacteria or fungi) and their susceptibility against antimicrobials enables the clinician to initiate an appropriate antimicrobial therapy and to guide diagnostic procedures. Whereas microbiological laboratory practice has been highly standardized, shortfalls in the preanalytic procedures in the ICU (indication, timing, volume, numbers, collection of blood cultures) have a significant effect on the diagnostic yield. Implementation strategies involving all ICU staff are needed to overcome the gap between recommended best practices and national guidelines. Finally, the BC frequency per 1,000 patient days should be established as a quality indicator in ICUs.