It is still the middle of winter, and for most of the readers of Shock, these are depressing times characterized by cold weather, dark skies, and snow, or even worse, my great nemesis, mud. It is the perfect time to enjoy the proverbial fire with a good journal, a glass of your favorite drink, and await the inevitable coming of Spring. If you are as fortunate as I, my reading material this week is the current issue of Shock. As I have commented frequently in the past, the age of electronic search engines, pdf files, and the now ubiquitous PMID or PMCID, the opportunity to peruse a journal is rapidly disappearing. I personally will find it a great loss because having a journal in one's hands is like going into a candy store, never really knowing what you will find, exploring new subjects and revisiting old ones.
This month, I decided rather than looking first for those articles directly relevant to my own field, I would read the journal cover to cover, and enjoy the delight of finding one article and then the next.
Anytime, “magnetic levitation” is in the title of an article in Shock, my curiosity is piqued and this is an excellent first read. Dr Andersen et al. (1) from the Beth Israel Hospital in Boston and Odense University in Denmark have used a novel technique combining magnetic levitation of blood leukocytes and microscopic phenotyping to distinguish dramatic differences between blood from healthy controls and septic individuals. Although the authors acknowledge the small sample sizes and the homogeneity of the patient population, this proof-of-principle report demonstrates that they might be on to something. Early stages of research are always full of optimism, but the findings are consistent with what we know regarding leukocyte morphology in the septic host. The option of having it bedside with rapid turnaround gives it a distinct advantage over flow cytometry, although alternative newer bedside technologies are being rapidly developed. Clearly, this is a rich environment that will require both validation and further technological development. Kudos to the research team.
Soussi et al. (2) from Paris and the PRONOBURN collaboration have provided an interesting and clinically important observation about the risk of mesenteric insufficiency in burn patients. Using a retrospective, case-control study design, the investigators demonstrate that low cardiac outputs and the presence of multiple organ failure are independent predictors of mesenteric ischemia, a deadly consequence of severe burns. There are two interesting take-home points: the mesenteric ischemia was nonocclusive suggesting that its etiology was primarily redistribution of blood flow and inadequate perfusion of the bowel, and secondary, acute mesenteric insufficiency was nearly uniformly lethal in burn patients. An excellent retrospective analysis suggesting that acute mesenteric insufficiency is a life-threatening complication of burn and should be closely monitored in patients with multiple organ dysfunction and low cardiac outputs. Easier to prevent than to treat, surely.
I am a big fan of using retrospective analyses in a discovery mode, obtaining preliminary data for a prospective trial. Clinical studies and randomized clinical trials in particular are a vast treasure of unplumbed observational data. Understanding their limitations still provides the opportunity to extract novel findings that can be tested moving forward. Here, Li et al. (3) from Chengdu, China, have conducted a retrospective cohort analysis of the neutrophil to lymphocyte ratio (NLR) in patients with acute respiratory distress syndrome (ARDS). NLR takes advantage of both the lymphopenia and granulocytosis associated with adaptive immune suppression and inflammation in ARDS patients. The results from 224 patients showed a clear relation between the NLR ratio at 24 h with mortality. NLR ratios at days 2 and 3 were also different between survivors and nonsurvivors, although the results were not as dramatic. Clearly, the studies demonstrate the value of an early NLR measurement, which importantly can be obtained routinely from a total and differential white blood cell count. Elevated NLRs at day 1 should be a clear warning signal to intensivists. My only wish had been that the authors had compared the predictive ability of NLR to other markers of both the severity of the ARDS, and the magnitude of the adaptive and innate immune systems. I am still unsure whether NLR is going to be more predictive than simply absolute lymphocyte count, procalcitonin, interleukin 6 (IL-6), or a plethora of other biomarkers.
Along the same lines, Laurikkala and the FINNRESCUI investigators conducted a retrospective analysis of 458 out of hospital cardiac arrest patients and used time-weighted lactate values as prognostic markers for subsequent outcome (4). Using regression models, the investigators demonstrated that both the time-weighted lactate values during ICU admission and the last lactate measurement from the ICU were independent predictors of 1-year survival. Surprisingly, neither admission lactate nor admission lactate clearance rates were good independent predictors of survival. Although this is not my strong area of knowledge, it is reassuring to know that admission lactates are not as important as the successful correction of tissue perfusion during ICU stay, and the degree of normality at discharge which are more important indicators of long-term success. Congratulation to the investigators on a potentially important observation, undoubtedly affected by the excellent ICU care provided in Finland.
Another example of important data extracted retrospectively from large data bases is the study of Yoshihiro et al. (5) in Hiroshima, Japan. Here, the investigators have extracted data from 1,180 patients at 42 academic institutions with severe sepsis and respiratory failure receiving recombinant human thrombomodulin (rhTM). Propensity score analyses were used to adjust the 356 subjects receiving rhTM and 824 subjects not receiving rhTM. Mortality was significantly reduced in the subjects receiving rhTM in patients with severe sepsis and respiratory failure, compared with subjects not receiving drug. The authors conclude that rhTM improves survival in this select cohort.
Retrospective studies like this in which subpopulations are culled from larger databases runs the risk of over-fitting data. There are a large number of past experiences where retrospective subgroup analyses have demonstrated efficacy only to have prospective trials fail to replicate the findings. One of the challenges with retrospective analyses is getting access to the data with sufficient granularity. For example, in this study, the amount of rhTM and the timing of the dosing were not known. Furthermore, there was no follow-up after hospital discharge. Although such data must be evaluated with a fair degree of skepticism, the findings here are large enough and significant enough to encourage a prospective, double blind, randomized controlled trial to answer the question definitively. Only through such rigorous study designs can these important questions be answered.
Dr Wade et al. in Houston have led the field for many years exploring the role of damaged glycocalyx and endothelial vesicles as endogenous alarmins and markers of endothelial injury. In this case-control study of trauma patients, Wade et al. (6) have examined a number of endothelial markers in 12 severely injured and 12 less severely injured trauma patients. A wide variety of analytes were compared between the two groups. There are a number of very interesting findings. Despite rather similar injury severity scores between the two groups, syndecan-1 concentrations differed 10-fold between subjects defined as having the endotheliopathy of trauma (EOT) and those that did not. Despite these dramatic differences in syndecan-1 concentrations, there were no differences in endothelial microvesicles irrespective of the antibody used. Plasma norepinephrine, sEselectin, sVE-cadherin, and histone-complexed DNA fragments levels were also similar. Only thrombomodulin, reductions in clot initiation, amplification, propagation and strength, and a greater frequency of transfusion, 92% versus 33% were seen in patients with EOT. The authors conclude that endothelial cellular damage or apoptosis does not differ, and, thus, endothelial glycocalyx disruption is the underlying primary cause of EOT. The subjects were initially categorized by their baseline syndecan-1 levels and not the severity of the injury per se, so it may be prudent to limit the conclusions to trauma patients with this degree of injury severity. In these patients, it seems reasonable to conclude that damage to the glycocalyx leads to EOT, but the role of microvesicles and cfDNA in severe trauma and EOT is still unresolved in patients with other injury severity.
Multiplex technologies for inflammatory cytokines and other mediators have revolutionized how we look at cytokine networks. These technologies generally use less sample and the reduced cost of analyte/sample has essentially eliminated the older ELISA technologies that measure one analyte at a time. Cytokine measurements have now become discovery tools with arrays reaching 40 to 50 cytokines measured simultaneously. In this report, Matsuura et al. (7) from Osaka have measured 11-cytokine panels in 38 burned subjects (>20 total body surface area) over periods of 1 month, and 12 healthy subjects. Not surprisingly, plasma IL-6, IL-8, IL-10, and MCP-1 concentrations were most predictive of survival, and these values were predictive early in the hospital course. The investigators argue that combining clinical markers with cytokine concentrations could be used to identify subjects at high risk of dying from their burn injuries.
Although these findings replicate a large number of earlier studies looking at cytokine concentrations in burn subjects, there are two novelties here. The first is that the large number of cytokines measured simultaneous focuses us on those that may be helpful, which are generally known, but also those that are not informative in terms of outcome. Second, they give us measures of relative value among the four cytokines of greatest predictive ability. Clearly, if the intent is to move this into the clinic with an Food and Drug Administration-cleared diagnostic, it would be important to identify the single cytokine that provides the most information. It would have been helpful if the authors had actually given us the area under the curves for the individual cytokines, and had conducted deLong's test to determine whether there were statistical differences among area under the receiver-operator curves. This latter point may not be necessary given the relatively small sample size and the need to validate these findings in larger cohorts, as discussed by the authors.
IL-33 is a member of the IL-1 superfamily and is thought to play a significant role in the polarization of the TH2 signaling pathway. ST2 is the presumed receptor. Much is known about IL-33's role in asthma, allergies, and contact dermatitis, but its role in T-cell polarization in burns is relatively unknown. Ruiz-Castilla et al. (8) in Barcelona studied 69 patients with burns ranging from 13% to 30% total body surface area and measured in the plasma a number of IL-33 superfamily members. Thirteen patients died of their burn injury. Surprisingly, neither IL-6, IL-8, nor IL-33 differed on day 1 after burn injury between survivors and nonsurvivors, while only sST2 differed. This difference held up at day 3. Using predictive modeling, the investigators report area under the receiver-operator curves of 0.73 and 0.85 for sST2, at days 1 and 3 post-burn. At day 3, elevated sST2 could best predict outcome even with cofounders.
There are a couple of issues that the authors need to consider. First, as they already discuss, the sample size is small and heterogenous. Those subjects that died were generally older with more severe burns. In this same issue of Shock, investigators from Japan have shown a significant value of IL-6 and IL-8 in predicting outcome in burns; yet here, the differences are minimal for IL-6 and IL-8. This suggests a difference in the two populations which only emphasizes the need for larger studies with well-defined burn populations. Clearly, the findings are evocative and require further study. Congratulations to the investigators for their work and their findings.
The role of the kidney in multiorgan failure secondary to septic or hemorrhagic shock has emphasized its importance in recovery or death. Acute kidney injury is one of the best predictors of outcome from severe sepsis or trauma. And yet, unless we are nephrologists, we think of the kidney as a homogenous organ whose response to shock is uniform. In this issue of Shock, Yan et al. (9) from the Netherlands investigated the inflammatory responses of three different renal microvascular segments, i.e., arterioles, glomeruli, and postcapillary venules, to hemorrhagic shock (HS) and resuscitation (HS/R) in mice, and to explore the effects of intervention with a nuclear factor-kappaB (NF-κB) inhibitor on these responses. Surprisingly, activation of the TIE-angiotensin network, and the cytokine cascade by HS/R affected different regions of the kidney, and blockade of NF-κB during the resuscitation phase only partly influenced the inflammatory response. This paper deserves a close read because it is a technological tour de force and its effort to focus on specific structural regions of the kidney. Although NF-κB blockade was not as effective as may have been desirable, the importance of the angiotensin and cytokine networks on different kidney regions is an outstanding observation in itself.
Moving from the kidney to the lung, Safavian et al. (10) from Toronto, Canada have examined the mechanisms driving M1 expansion in the lungs of mice subjected to HS/R. Their straightforward hypothesis is that HS/R increases the lung responsiveness to secondary injury by increasing the M1/M2 ratio and increasing the inflammatory response. Here, they phenotyped lavage fluid from mice undergoing HS/R with or without intratracheal endotoxin challenge 2 hours after resuscitation. Some animals received intratracheal administration of M2 polarized macrophages. The findings are somewhat surprising in that the increased M1/M2 ratio was predominantly due to a loss of M2 cells rather than expansion of M1 subgroups. Readministration of M2 macrophages attenuated the inflammatory response in the lung to lipopolysaccharide. The novelty here is the observation that therapies should be perhaps targeting expansion of the M2 population after hemorrhagic shock and resuscitation, rather than attenuating M1 phenotypes. Another outstanding and evocative paper from the Toronto group.
Sticking with the lung for a few minutes, Gugliandolo et al. (11) from Messina, Italy, have examined the role of the TLR4 signaling pathways and the inflammasome, and particularly, PPARα in the host response to Pseudomonas pneumonia. Here, the investigators used a genetic approach, employing TLR4 and PPARα null mice. Not surprisingly, Pseudomonas pneumonia was more lethal in both TLR4 and PPARα null mice, indicating an essential role for both these proteins in host protective immunity. Interestingly, both pathways are critical for the magnitude of the inflammatory response either directly through NF-κB or through processing of IL-1β, IL-18. In both cases, NF-κB translocation to the nucleus was increased in all of the knockout mice, and although this may appear superficially contradictory, it is more consistent with the increased pathogenicity, lung damage, and death in the null mice. My take from these important studies is that TLR4 and the inflammasome signaling pathways both activate NF-κB-dependent early response genes essential for host protective immunity, and loss of either increases mortality to Pseudomonas pneumonia. Mortality is secondary to an exaggerated inflammatory response even in the absence of TLR4 and PPARα.
Probably one of the most exciting focus of sepsis research in 2018 to 2019 has been on how sepsis influences the brain. Neuroinflammation in sepsis is now being reported and is being observed for months after resolution of the initial inflammatory event. Long-term cognitive and behavioral changes are being recognized much more frequently. Yet, little is still known about how the brain responds to septic insults. Here, Kurtz et al. (12) from Rio De Janeiro conducted brain microdialysis in pigs subjected to the hyperdynamic phase of sepsis. In this porcine model of resuscitated sepsis, the investigators found increased oxidative stress, exacerbated activity of the glutamate/glutamine cycle and increased glucose utilization by the brain, however without any evidence of decompensated energy metabolism. Whether this increased oxidative stress is responsible for astrocyte and glial cell activation, commonly seen in sepsis survivors needs to further elucidated.
Back to the lung and therapeutic interventions aimed at reducing lung injury, Wang et al. (13) from Shanghai examined the role of tissue factor pathway inhibitor (TFPI) on endotoxin induced lung injury produced by endotoxin administration. I have to disclose my bias as I participated in both the preclinical baboon studies with recombinant TFPI as well as the clinical trials in sepsis. Here, the investigators took the alternative approach by removing TFPI from endothelial tissues by using Tek-TFPI floxed mice. The striking findings seen with this powerful genetic tool were how pervasive the importance of TFPI is in modulating the inflammatory response to endotoxin. These endothelial TFPI null mice were extremely sensitive to lipopolysaccharide administered into the lung. TFPI deletion markedly exacerbated histopathological changes in lung, and changes in protein and fluid extravasation, proinflammatory cytokines TNFα, IL-1β, and IL-6 in broncho-alveolar lavage fluid in lung in response to endotoxin. The number and infiltration of white blood cells from lavage fluid and lung tissue of TFPI null mice with acute lung injury was increased compared with wild-type mice challenged with endotoxin. Clearly, endogenous TFPI plays an essential role in modulating lung injury. I only wish that pharmacologic levels of TFPI could have had the same magnitude of response in patients with sepsis, with either normal or depleted TFPI levels.
This next manuscript demonstrates the power of collaboration and the use of synergistic animal models. Investigators from Germany, Portugal, Switzerland, and Italy have conducted a two-species study of specific and nonspecific potassium channel inhibitors in an ovine model of endotoxic shock and rat model of septic shock (14). These are both excellent randomized trials, one focusing on macrocirculatory effects, the other microcirculatory effects. The power of this combined approach is the ability to ascertain simultaneously both potentially beneficial and adverse effects of nonspecific versus specific potassium channel inhibitors. Selective inhibition of potassium adenosine triphosphate-channels in ovine endotoxemic shock with specific inhibitors partially restored vasomotor tone without exerting harmful effects on intestinal microcirculation in septic shock in rats. On the contrary, nonselective potassium-channel inhibition showed deleterious effects in both models, including impaired microcirculation and decreased survival time. Importantly, these findings could not have been reproduced by focusing on a single animal model. Future research on selective potassium channel inhibitors in vasodilatory shock appears to be warranted.
Gu et al. (15) from Changsha, and their collaborators from Pittsburgh and Manhasset have made a novel observation regarding a previously unknown role for the alarmin receptor, TLR4. TLR4 differs from other receptors of the TLR signaling family in that ligands for the TLR4 receptor can signal both through cytosolic MyD88 pathways directly to NF-κB activation, and through cytosolic TRIF pathways directed at type I interferon production. Here, the authors convincingly demonstrate that microbial outer membrane vesicles (OMV) can bind to TLR4 and be transported cytosolically by TRIF, and their transport is dependent upon both the production of type I interferons and the guanylate-binding proteins. Interestingly, the investigators provide a clearly-described mechanism by which bacterial products contained in OMV can activate both plasma bound TLR4 receptors and cytosolic inflammasomes. Importantly, it also provides an explanation for both the redundant importance of bacterial recognition and an explanation for the exaggerated inflammatory response driven by both TLR signaling and the inflammasome. It also clarifies the somewhat unexplained observation that inflammasome activation is markedly increased when preceded by a TLR activation. Hopefully, this will be a highly read and cited paper. It deserves as much for the novelty of the observation and the convincing nature of the experimental data.
Finally, one would be doing themselves a significant disfavor if they did not read the Letter to the Editor by Leitje and his collaborators (16), and the response by Jansen et al. (17). The discussion is an important one and is focused on the role of cell-free mitochondrial and nuclear DNA as alarmins, resulting from cell injury and death. In this particular case, the argument centers around the role of mitochondrial DNA in producing the kidney injury associated with sepsis. In general, the argument is more broad. Are cell-free DNA recognized by alarmins through signaling pathways shared by microbial products and do they produce the tissue and organ injury that we see in sepsis, trauma, operative injury. Or, are they merely biomarkers reflecting the magnitude of endogenous tissue injury. It is a spirited discussion, no consensus is achieved, and the final results will require additional investigation.
1. Andersen MS, Lu S, Lopez GJ, Lassen AT, Shapiro NI, Ghiran IC. A novel implementation of magnetic levitation to quantify leukocyte size, morphology, and magnetic properties to identify patients with sepsis. Shock
2. Soussi S, Taccori M, De Tymowski C, Depret F, Chaussard M, Fratani A, Jully M, Cupaciu A, Feerry A, Benyamina M, et al. for the PRONOBURN group. Risk factors for acute mesenteric ischemia in critically ill burns patients—a matched case-control study. Shock
3. Li W, Ai X, Ni Y, Ye Z, Liang Z. The association between the neutrophil-to-lymphocyte ratio and mortality in patients with acute respiratory distress syndrome: a retrospective cohort study. Shock
4. Laurikkala J, Skrifvars MB, Bäcklund M, Tiainen M, Bendel S, Karhu J, Varpula T, Vaahersalo J, Pettilä V, Wilkman E. the FINNRESUSCI study group. Early lactate values after out-of-hospital cardiac arrest: associations with one-year outcome. Shock
5. Yoshihiro S, Sakuraya M, Hayakawa M, Ono K, Hirata A, Takaba A, Kawamura N, Tsutsui T, Yoshida K, Hashimoto Y. Recombinant human-soluble thrombomodulin contributes to reduced mortality in sepsis patients with severe respiratory failure: a retrospective observational study using a multicenter dataset. Shock
6. Wade CE, Matijevic N, Wang Y-WW, Rodriguez EG, Lopez E, Ostrowski SR, Cardenas JC, Baer LA, Chen T-A, Tomasek JS, et al. Absences of endothelial microvesicle changes in the presence of the endotheliopathy of trauma. Shock
7. Matsuura H, Matsumoto H, Osuka A, Ogura H, Shimizu K, Kang S, Tanaka T, Ueyama M, Shimazu T. Clinical importance of a cytokine network in major burns. Shock
8. Ruiz-Castilla M, Bosacoma P, Dos Santos B, Baena J, Guilabert P, Marin-Corral J, Mascians JR, Roca O, Barret JP. Soluble suppression of tumorigenicity-2 predicts hospital mortality in burn patients: an observational prospective cohort pilot study. Shock
9. Yan R, van Meurs M, Popa ER, Li R, Zwiers PJ, Zijlstra JG, Moser J, Molema G. Early heterogenic response of renal microvasculature to hemorrhagic shock/resuscitation and the influence of NF-κB pathway blockade. Shock
10. Safavian D, Leung CH, Kapus A, Ailenberg M, Szaszi K, Shani R, Di Ciano-Oliveira C, Ghazarian M, Rotstein O. Hemorrhagic shock/resuscitation reduces the M2 phenotype of alveolar macrophages: a potential mechanism contributing to increased LPS-induced lung injury. Shock
11. Gugliandolo E, Fusco R, Ginestra G, D’amico R, Bisignano C, Mandalari G, Cuzzocrea S, Di Paola R. Involvement of TLR4 and PPAR-α receptors in host response and NLRP3 inflammasome activation, against pulmonary infection with pseudomosas aeruginosa. Shock
12. Kurtz P, d’Avila JC, Prado D, Madeira C, Vargas-Lopes C, Panizzutti R, Azevedo LCP, Bozza FA. Cerebral multimodal monitoring in sepsis: an experimental study. Shock
13. Wang BQ, Shi M, Zhang JP, Wu X, Chang MJ, Chen ZH, Shen HH, Song YL, Zhou J, Bai CX. Knockdown of TFPI-anchored endothelial cells exacerbates lipopolysaccharide-induced acute lung injury via NF-κB signaling pathway. Shock
14. Hessler M, Pinto BB, Arnemann P-H, Kampmeier T-G, Seidel L, Morelli A, Van Aken H, Westphal M, Rehberg S, Ertmer C. Differential effects of selective and nonselective potassium channel inhibitors in ovine endotoxemic shock (macrocirculation) and in a rat model of septic shock (microcirculation). Shock
15. Gu L, Meng R, Tang Y, Zhao K, Liang F, Zhang R, Xue Q, Chen F, Xiao X, Wang H, et al. Toll-like receptor 4 signaling licenses the cytosolic transport of lipopolysaccharide from bacterial outer membrane vesicles. Shock
16. Letter to the Editor: Leijte GP, Pickkers P, Kox M. Mitochondrial DNA: Innocent in plasma, but guilty in urine? Shock
17. Reply to the Letter to the Editor: Jansen MPB, Roelofs JJTH, Leemans JC. Mitochondrial DNA: Innocent in Plasma, but Guilty in Urine? Shock