Setting the baseline to fight Gram-negative bacteraemia: the necessity of epidemiological insights

Nosocomial Gram-negative bacteraemia is a frequent infectious complication associated with increased morbidity and, possibly, mortality [1–3]. Gram-negative pathogens favour immunosuppressed patients or otherwise patients with a debilitated physical condition. Unfortunately, over the past decades, the proportion of patients with suboptimal immune defence increased steadily due to the widespread use of immunosuppressive agents for patients with conditions such as chronic obstructive pulmonary disease or solid organ transplantation. In addition, improved survival rates in chronic diseases has led to a growing pool of patients at high risk for infectious complications due to underlying disorders with a down-regulated immune function (e.g. renal failure or hepatic cirrhosis) [4,5].

In the 1960s and 1970s, Gram-negative bacteraemia was considered a bad omen indicating imminent death and several researchers even questioned whether antibiotic therapy was capable of altering the outcome [6–9]. Since the 2000s, it became clear that the high mortality associated with Gram-negative bacteraemia was mainly due to the severity of underlying disease and acute illness and that – in the presence of prompt and appropriate empiric antibiotic therapy – the attributable mortality could be reduced to non-significant proportions [10,11]. At the same time, the insight grew that a substantial proportion of healthcare-associated infections could be prevented and that these complications could no longer be considered unavoidable even in the presence of severe illness [12,13]. As such, Gram-negative bacteraemia poses a clinical challenge in prevention and management.

In this issue of Infectious Diseases, Delle Rose et al. describes a detailed study of nosocomial bloodstream infection caused by Gram-negative bacteria [14]. The report mentions data on incidence, source of the bacteraemia, species distribution and multidrug resistance, rates of appropriate empiric antibiotic therapy and outcome. Such in-depth epidemiological analyses serve the clinician in the task to prevent and treat bacteraemia. In fact, the data set a baseline that is indispensable to adequately define objectives in quality and safety projects.

First, Delle Rose et al. report an overall incidence of 12.8 Gram-negative bacteraemia per 1000 hospital admissions. Comparison with other institutions is difficult as this metric is particularly case-mix dependent. However, incidence is an essential baseline to monitor the overall performance of infection prevention. An important aspect that is often neglected regarding the monitoring of the incidence of bacteraemia is the protocol (if any) to sample blood cultures, more precisely the indications for sampling the cultures.

Protocols often exclusively define a fever threshold as an indication for sampling blood cultures (e.g. >38.0°), thereby passing by on hypothermia, or sudden confusion as signs of bacteraemia. The latter parameter is of particular value in elderly who have a blunted inflammatory response allowing sepsis to occur without fever. A blood culture protocol defining a larger set of indications for sampling blood cultures might lead to the detection of a greater burden of the disease. Of course, in this regard, sufficient attention must go out to the risk of blood culture contamination due to a poor aseptic sampling technique.

Second, insights in the source of bacteraemia can be of value to put emphasis on prevention of particular primary infections, such as pneumonia, catheter-associated urinary tract infection or central line-associated bloodstream infection. Targeted quality improvement initiatives proved highly successful in preventing a variety of infection, including hospital-acquired bacteraemia and central line-associated bacteraemia [15–21]. Concerning the prevention of bacteraemia, the use of chlorhexidine-impregnated washcloths has become trendy in critical care units and general wards. Of note however, chlorhexidine washcloths appear to be more effective in reducing Gram-positive rather than Gram-negative bacteraemia [22].

Third, thorough epidemiological study comes with knowledge of local microbial ecology, that is species distribution and involvement of multidrug resistance. Evidently this information is crucial to evaluate effectiveness in infection prevention and control. Infection rates need to be evaluated over large periods of time (i.e. at least quarterly) as the epidemiology of infection is prone to natural fluctuations as it depends on import of multidrug resistant clones (patients positive for ESBL on admission) and adherence with prevention recommendations [23–26].

Microbiological insights are also useful to optimize the rate of empiric appropriate antibiotic therapy. While regional or hospital-wide data serve for the development of (inter-) national recommendations or local protocols [27], information about the individual colonization status is highly valuable to tailor empiric antimicrobial therapy for the index patient [28–30]. Local microbiological insights have become close to indispensable for steering empiric therapy since classic risk factors for multidrug resistance, such as length of hospitalisation and prior antibiotic exposure, have lost their predictive value [31]. Indeed, as multidrug resistance becomes more and more an issue of the community rather than an exclusive hospital problem, the need for patient-specific data to guide antimicrobial therapy has become increasingly important.

Besides appropriate empiric antibiotic therapy the likelihood of survival of a patient with Gram-negative bacteraemia depends on age, associated acute organ derangements, underlying conditions and immune status, the pathogen involved and its resistance pattern. All these aspects are unchangeable, except antibiotic therapy. High rates of appropriate antibiotic therapy initiated within the critical time frame of the first 24–48 hours have repeatedly been demonstrated as associated with better survival rates [32–34].

However, the favourable effect of early appropriate therapy depends on other associated prognostic factors. The study by Delle Rose et al. elegantly describes the impact of appropriate therapy related to the extent of associated organ failures as evidenced by the SOFA scores [14]. For example, the difference in predicted mortality between adequate and inadequate antibiotic therapy for a patient with a SOFA of 10 is approximately 28% (32% vs. 60% in patients with respectively adequate and inadequate antibiotic therapy). Towards the extreme of disease severity, that is a SOFA score >15, the beneficial effect of adequate therapy shrinks indicating that in these patients’ overall disease severity is more determinative for the outcome. Therefore, severity of acute illness must be considered as well when evaluating mortality rates associated with Gram-negative bacteraemia. One can assume that prevention will rather be effective in patients with mild or moderate disease severity. This implies that general disease severity in patients in which prevention failed will be higher, with a less favourable impact of adequate antibiotic therapy.

The data reported by Delle Rose et al. are exemplary for how basic epidemiology of nosocomial infection might constitute a solid baseline for designing and evaluating quality improvement interventions targeting prevention as well as management of Gram-negative bacteraemia.

Disclosure statement

No potential conflict of interest was reported by the author.