Sepsis is a major cause of death worldwide, not least because complex interventions need to be provided within a short window of opportunity. Evidence-based guidelines for the treatment of sepsis are therefore welcome, providing a common ground for all clinicians involved in decision-making regardless of their expertise. Such guidelines should therefore serve as an overarching reference document. As previously stated, ‘Guidelines are the product of an explicit, systematic approach to the evaluation and synthesis of available information on a particular clinical topic. They are not a compilation of truths, but are a summary of what is accepted by the authors as the best available evidence at that time’ [1].

As the evidence base evolves over time, the new Surviving Sepsis Campaign (SSC) guidelines [2] are timely and important. We appreciate the tremendous amount of time invested by the experts who formulated the new guidelines to provide the intensive care community with a clear and comprehensive manuscript. We also recognize the challenge of providing evidence-based guidelines when the strength of evidence available to direct recommendations is limited. Indeed, the large majority of randomized, controlled trials (RCTs) performed over the last three decades in intensive care medicine, including those in sepsis, have shown no significant beneficial effect of the tested intervention on outcomes [3]. At face value, this may simply suggest that the myriad of interventions that have been tested are all ineffective. However, it is more likely that subsets of patients who benefit from specific treatments have yet to be identified. The often broad patient inclusion criteria could easily lead to dilution of positive findings by non-responders, or to positive effects in some patients being offset by harm in others [4]. This treatment effect heterogeneity has been clearly indicated in many trials (for example [5,6,7,8]).

The most severely ill patients usually suffer less from a single condition than from a complex physiological imbalance that defies specific disease definitions. Management that strictly adheres to guidelines may not necessarily be the best option. However, personalizing management of these patients mandates that the treating clinician appreciates the clinical implications of the underlying disease(s) and individual host factors (chronic health status, physiology, and physiological reserves). Additionally, the individual response to interventions must be closely monitored and, depending on that response and whether or not the pre-set goal of that intervention has been achieved, the clinician can then decide to maintain or alter the intervention accordingly. We fully acknowledge this is demanding in resources, time, personnel, and bedside expertise. It is more convenient and less labor-intensive to adopt a generalized approach, yet customization is surely the underlying foundation of intensive care. When we forget the fundamental importance of individualized management, the ultimate result is that few interventions improve meaningful patient outcomes, especially mortality [3]. We must acknowledge and embrace natural (patho) physiological variability; sepsis is no exception to this rule. Sepsis encompasses a huge spectrum of clinical situations in terms of the type of patient involved, the clinical presentation and the response to treatment. This variability should serve to direct clinical management, enabling the clinician to adapt “recommended” care according to the specific needs of that patient. As David Sackett, the father of evidence-based medicine (EBM) stated, “Good doctors use both individual clinical expertise and the best available external evidence, and neither alone is enough. Without clinical expertise, practice risks becoming tyrannised by evidence, for even excellent external evidence may be inapplicable to or inappropriate for an individual patient. Without current best evidence, practice risks becoming rapidly out-of-date, to the detriment of patients.” [9].

The conundrum is how to encourage guideline implementation yet at the same time promote personalized medicine. The new guidelines [2] offer little leeway for adapting the recommendations to the idiosyncrasies of each and every patient. Many of the recommendations attempt to fit most (if not all) of our patients. Sets of recommendations have been translated into bundles of care to be applied within a rigid time-frame to the “average” patient. This approach has sacrificed precision for homogenization and expediency. Some allowance for breaking the “one size fits all” guideline mold that has taken root in the last two decades would have been a daring but welcome and timely change.

We have therefore taken the liberty of putting forward some proposals of how guidelines could be adapted to individualize care (Table 1). We fully acknowledge that these recommendations are primarily based on pathophysiological considerations and clinical experience rather than on RCT data.

Table 1 Twenty recommendations to individualize interventions in the early resuscitation of patients with sepsis
Full size table

The way forward is to be bold enough to question how we can do better. It may be time to move from a mass approach, based on pragmatic studies performed on heterogeneous populations, to tailored studies allowing both dissection and integration of the information collected in specific sepsis phenotypes. We need to embrace theragnostic approaches, using biomarkers to identify only those patients for whom the intervention is suited, and to titrate dose and duration to optimal effect. Crucially, we need to restore the application of physiology and biochemistry to the forefront of our clinical practice. Returning to Sackett: “Evidence based medicine is not "cookbook" medicine. (…) External clinical evidence can inform, but can never replace, individual clinical expertise, and it is this expertise that decides whether the external evidence applies to the individual patient at all and, if so, how it should be integrated into a clinical decision.” [9]. Sepsis is clearly one instance where “one size does not fit all”.