现在的位置: 首页研究点评, 进展交流>正文
[Lancet Respir Med最新论文]:有关全身性感染临床试验的再思考
2017年08月10日 研究点评, 进展交流 暂无评论

Clinical trial research in focus: rethinking trials in sepsis

David Grimaldi, Jean-Louis Vincent

Lancet Respir Med 2017; 8: 610-611

Published: August 2017

DOI: http://dx.doi.org/10.1016/S2213-2600(17)30268-0

Sepsis is a frequent condition, defined as life-threatening organ dysfunction associated with a severe infection.1 Over the past three decades, progress in understanding of the pathophysiology of sepsis has led to the development of a large number of potential new therapies for sepsis. Many of these underwent preclinical testing with promising results and moved into clinical trials with the hope that they would improve survival from sepsis. Some of these interventions successfully passed early phase trials and underwent testing in large multicentre randomised controlled trials. However, here the success story stopped, with none of the agents successfully shown to improve outcomes. Sepsis treatment is, therefore, still largely based on antibiotic therapy, source control, and haemodynamic resuscitation.2 The fact that so many attempts to improve sepsis outcomes have failed must prompt a rethink of clinical trial design and conduct in this field. We will detail some of the reasons why so many trials were negative and make some suggestions as to how to move forward.

It is important to remember that sepsis is not a disease, but a heterogeneous syndrome associated with a specific infectious disease. It can be caused by a wide range of different pathogens, acting on different organs, resulting in an interplay of diverse pathological processes. Genetic variability exists in host and pathogen, and comorbidities and the timeframe of the infection can also affect disease presentation. These different aspects were included in the PIRO acronym,3 with P for predisposition (factors present before sepsis; eg, genetic factors, comorbidities, underlying immunosuppression, alcoholism), I for infection (microorganism, site of infection), R for response (degree and type of host response, including levels of procalcitonin and other markers), and O for organ dysfunction (patterns and severity of organs affected). These multiple factors and even some stochastic effects explain the diversity of the sepsis phenotype. This heterogeneity generates a strong noise-to-signal ratio, making any demonstration of efficacy very difficult. Moreover, in view of the great diversity, it is not too surprising that one intervention, as tested in most of the clinical trials, will not be effective in all patients with sepsis.

Clinical trials in sepsis have generally used mortality as the primary study endpoint, including, more recently, longer follow-up periods of 90 days. However, mortality is influenced by many factors, including the primary disease, comorbidities, non-sepsis-related events, and especially end-of-life decisions, which can make it less of a clear-cut endpoint. In oncology, progression-free survival has replaced crude mortality rates as the primary endpoint in many studies. In the intensive-care unit (ICU) setting, endpoints of morbidity, such as organ dysfunction or ICU-acquired infection, could be used, selected according to the expected effect of the intervention under study. Increasingly, studies are incorporating such endpoints into their design.4 Mortality should not be totally discarded as an endpoint, but could be used in combination with morbidity or as a safety variable rather than an endpoint per se.

In addition to the issue of heterogeneity, the suitability of some of the proposed treatments is questionable. Nearly all had promising results in preclinical studies, mostly using rodent models of sepsis. However, these models have some important general limitations, including their lack of similarity to the clinical reality, generally using previously healthy animals without comorbidities and often administering the intervention before or very soon after sepsis initiation. The short-term observation periods are also a limitation. There are also specific limitations for different sepsis models. For example, endotoxin challenge reproduces only in part the complex process created by living microorganisms.5 Similarly, models of peritonitis induced by caecal ligation and puncture have quite variable outcomes, and pneumonia and infusion of live bacteria do not always result in sepsis. Finally, rodents have a different immune system to that of human beings. The most striking example of this is that meningococci are unable to infect non-human mammals, whereas they are the cause of the most severe form of sepsis in patients. More generally, it has been observed that the transcriptional response of circulating leucocytes of mice and human beings share only 50% similarity.6 Findings from basic research should, therefore, be interpreted cautiously before starting randomised trials, and every effort should be made to do translational studies.

Targeting of the immune response is an appealing strategy. The role of proinflammatory mediators in sepsis pathogenesis was the first to be identified and the search for therapeutic agents largely focused on anti-inflammatory interventions. But attempts to modulate the proinflammatory response, although largely successful in animal models, were disappointing in patients. Indeed, it has become increasingly obvious that sepsis also induces a long-lasting anti-inflammatory response in many patients resulting in the development of acquired immunosuppression, thus leading to the concept of immune dysregulation.7 In some patients, the systemic proinflammatory response can be so brief that many patients with sepsis are first examined when they already have a predominantly anti-inflammatory response.8

Targeting of sepsis-induced immune suppression might prevent secondary infection. Several molecules are currently being tested in phase 2 and 3 trials aimed at reversing monocyte deactivation (interferon gamma, granulocyte/macrophage-colony stimulating factor), inhibiting sepsis-induced apoptosis (interleukin 7, interleukin 15) or targeting the immune checkpoint (anti-programmed cell death-1/programmed cell death ligand-1 antibodies).9 Importantly, these ongoing studies will focus on patients with features of immune suppression (eg, low monocyte HLA-DR or lymphopenia). A key shortcoming of such strategies is that circulating cells might not reflect the immune status of non-circulating cells. Moreover, different components of the immune system can be activated and inhibited at the same time.

In conclusion, sepsis should not be considered as a disease, but rather as a complex disorder that calls for the development and application of personalised medicine. The specificity of the host–pathogen interaction should encourage a cautious approach when interpreting the results of sepsis therapies in animal models and translating them to patients. Trials should be done in well defined groups of patients believed to be likely to benefit from the intervention being tested. Biomarkers can help in this task and a biomarker-driven approach to clinical trials is beginning to be used in sepsis. We should also accept that mortality is not the sole endpoint of value and that improvement in organ failure or prevention of ICU-acquired infections is a substantial improvement in patient management. In the future, we will no longer focus on so-called sepsis drugs, but rather on specific interventions in patients with a particular sepsis phenotype,10 enabling the right treatment to be given to the right patient at the right time.

David Grimaldi, *Jean-Louis Vincent 

Department of Intensive Care, Erasme Hospital, Université Libre de Bruxelles, 1070 Brussels, Belgium jlvincent@intensive.org

We declare no competing interests.


1 Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis and septic shock (sepsis-3). JAMA 2016; 315: 801–10.

2 Rhodes A, Evans LE, Alhazzani W, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock: 2016. Crit Care Med 2017; 45: 486–552.

3 Levy MM, Fink MP, Marshall JC, et al. 2001 SCCM/ESICM/ACCP/ATS/SIS international sepsis definitions conference. Crit Care Med 2003; 31: 1250–56.

4 de Grooth HJ, Geenen IL, Girbes AR, Vincent JL, Parienti JJ, Oudemans-van Straaten HM. SOFA and mortality endpoints in randomized controlled trials: a systematic review and meta-regression analysis. Crit Care 2017; 21: 38.

5 Eisenstein TK, Deakins LW, Killar L, Saluk PH, Sultzer BM. Dissociation of innate susceptibility to Salmonella infection and endotoxin responsiveness in C3HeB/FeJ mice and other strains in the C3H lineage. Infect Immun 1982; 36: 696–703.

6 Seok J, Warren HS, Cuenca AG, et al. Genomic responses in mouse models poorly mimic human inflammatory diseases. Proc Natl Acad Sci USA 2013; 110: 3507–12.

7 Boomer JS, To K, Chang KC, et al. Immunosuppression in patients who die of sepsis and multiple organ failure. JAMA 2011; 306: 2594–605.

8 Davenport EE, Burnham KL, Radhakrishnan J, et al. Genomic landscape of the individual host response and outcomes in severe sepsis. Lancet Respir Med 2016; 4: 259–71.

9 Hotchkiss RS, Moldawer LL, Opal SM, Reinhart K, Turnbull IR, Vincent JL. Sepsis and septic shock. Nat Rev Dis Primers 2016; 2: 16045.

10 Vincent JL. Individual gene expression and personalised medicine in sepsis. Lancet Respir Med 2016; 4: 242–43.


您必须 [ 登录 ] 才能发表留言!