Underlying assumptions of sepsis treatment

Why have clinical trials in sepsis failed? One potential answer is that the underlying assumptions about immune function in sepsis are mistaken.

What are the underlying assumptions?

The main assumption of immunomodulatory treatments for sepsis is that the observed host response is different from the optimal host response.

Is it?

We have good reason to think that any individual’s measured cytokine or chemokine  (e.g. IL-1β or TNF-α)  is probably not optimal.


Because significant inter-individual variation exists for every trait. Logic dictates that  there is no way that all of the different measurements that show variation in individuals in sepsis can be simultaneously optimal.


For most traits, it is unlikely that the non-optimality occurs is the same direction for all individuals.

Therefore, a second underlying assumption for sepsis immune treatments is that we know the direction of the non-optimality (usually involving an excessive inflammatory response). We also assume that patients will do better if we intervene in the opposite direction.

A third assumption is that intervening on a single biomarker will adjust host immune function in a way that increases fitness (in the biological sense). What if it doesn’t?

One fact that undermines the last assumption is that immunity relies on highly complex and interlinked regulation, characterized by compensation, and redundancy. In other words, interrupting TNFα function is likely to produce compensatory changes in other cytokines that might produce an equivalent functional result.

When we use medications in sepsis, are the compensatory effects and other pleiotropic changes in gene expression optimal or non-optimal? We don’t know.

Moreover, there is reason to believe that immune regulation is resistant to interference with any single immune signal.


Many pathogens have evolved ways to interfere with immune functions. These manipulations allow pathogens to establish an infection. One example is the inhibition of NF-kb and IL-8 in Pseudomonas aeruginosa pneumonia. Also read here and here and here (for more examples and a list of relevant references)

Host pathogen arms races have resulted in adaptations enhancing virulence in pathogens, and counter-adaptations in their hosts. An excellent example of this was observed in the iron-sequestering protein transferrin and Neisseira meningitidis, recently described in the New York Times.

Host-pathogen arms races might be responsible for immune system robustness that is resistant to inhibition of any one immune mediator. Hence, we observe complexity and redundancy in immune regulation.

The outcome of these co-evolutionary dynamics might yield functional results far closer to optimality than any single immune mediator. It would also create conditions resistant to “improvement” by blocking the endogenous host response. This last part, blocking the endogenous host response, typifies every failed anti-inflammatory treatment tried in humans thus far.

© Joe Alcock MD 2014

Sepsis – what have we learned?

Sepsis is an important cause of mortality, causing an estimated 60,000 deaths yearly. Sepsis is also expensive to treat, and is associated with expensive medical procedures, such as life support and intensive care. Despite advances in supportive care over the last 30 years, the mortality rates have remained stubbornly high, as much as 30-40%.

Sepsis Mortality

In septic shock, treatment includes early and appropriate empiric antibiotics, source control, and supportive care. Of these, antibiotics and source control are thought to be the most important, but surprisingly little data support the elements of care. Controversy surrounds most aspects: fluids, pressors, oxygen replacement, and enhancement of cardiac output, to name a few.  Kox et al wrote an illuminating recent review on aggressive therapy in sepsis and critical illness entitled Less is more in critically ill patients. In light of these ongoing controversies, can an evolutionary perspective of disease provide any guidance?

One example involves fluid administration:

Recent research in sepsis has also hinted at a phenomenon thus far not even considered—that hypotension is an adaptive response, honed over millions of years of evolution, which may mean that we are actually making patients worse when we thought we were helping them.“-David Anderson

Read Anderson’s account of controversy in giving fluid therapy for sepsis and critical illness.

And: Conservative Fluid Management In Intensive Care Medicine by David Gattas

A second, related possibility is that the immune systems response is also adaptive in septic shock. If so, anti-inflammatory treatments won’t work. Supporting that notion is the recent clinical experience with recombinant activated protein C:

Persistently high sepsis mortality prompted many researchers and funding organizations to seek anti-inflammatory treatment to decrease the apparently harmful immune effects of sepsis, which has been described as an out of control inflammatory response.

One thing that biomedical researchers noticed was that septic patients who die often had low levels of activated protein C.  A recombinant form of activated protein C, called  Xigris, has anti-inflammatory and anticoagulant properties. Since inflammation was thought to be out of control in the systemic inflammatory immune response, it stood to reason that an inhibitor of inflammation and clotting might reduce deaths in sepsis.

In 2001 the PROWESS study appeared to show just that.

10 years late, on October 25, 2011 the FDA recommended that Xigris be withdrawn from the market. Remarkably, the company selling Xigris, Eli Lilly, was able to profit from this ineffective drug for nearly the entirety of the time Xigris was under patent. The long road to arrive at this conclusion is an interesting story:

Epitaph for Xigris – I never worked a day in my life!

What lesson from Xigris’s failure can we learn about the functioning of the immune system, about sepsis, and evolution? Notably, Xigris is not the only high-profile failure for sepsis immunomodulatory therapy. A more recent anti-inflammatory therapy, a toll-like receptor TLR-4 blocker known as Eritoran, also failed spectacularly in clinical trials.

Not be deterred, a new therapy for sepsis targeting the pro-inflammatory cytokine TNF-alpha, based on a compound called CytoFab, was recently tested. How do you suppose that one went? Shockingly, it too failed.

Why have clinical trials in sepsis failed?

(Mostly) failed trails of immunomodulators in sepsis

Table 1 showing failed clinical trials of immune and other therapies in sepsis above, and the abstract below, is reproduced from Marshall Why have clinical trials in sepsis failed? Trends Mol Med 2014

“The systemic inflammatory response is biologically complex, redundant, and activated by both infectious and noninfectious triggers. Its manipulation can cause both benefit and harm. More than 100 randomized clinical trials have tested the hypothesis that modulating the septic response to infection can improve survival. With one short-lived exception, none of these has resulted in new treatments. The current challenge for sepsis research lies in a failure of concept and reluctance to abandon a demonstrably ineffectual research model. Future success will necessitate large studies of clinical and biochemical epidemiology to understand the course of illness, better integration of basic and clinical science, and the creation of stratification systems to target treatment towards those who are most likely to benefit.”

These examples suggest that continuing to search for elusive so-called “magic bullets” in sepsis is a losing strategy. As cited in a recent article about the failure of anti-TNF-α in sepsis:  “Success is the ability to go from one failure to another with no loss of enthusiasm,” a quotation attributed to Sir Winston Churchill. On the other hand, repeating same thing while expecting different results is also…well you know the cliché.

Student writing assignment:

You have a relative living in Amsterdam who was approached by a Dutch researcher about enrolling in a real study of an anti-inflammatory agent in endotoxemia: The study, entitled in vivo effects of C1-esterase inhibitor on the innate immune response during human endotoxemia is underway. What do you advise her to do and why?

Late Update: Another area of active investigation is the possible beneficial effect of the lipid-lowering drugs known as statins on sepsis and its complications. The idea is that in addition to reducing cholesterol, statins have anti-inflammatory and anti-coagulant effects. (Sounds a lot like Xigris, doesn’t it.)

The blog Life in the Fast Lane has published a nice summary of what is currently known about statins in sepsis that is absolutely worth a read. Germane to our discussion, NEJM published a recent test of statins in the sepsis-associated lung disease. Result: Failed.  Last year, a randomized trial of atorvostatin in sepsis failed to show an improvement in 90 day mortality. The authors of that study hold out hope that statins will be proven efficacious in sepsis, and the final word on this topic has not been published. Another randomized trial of statins in sepsis is currently recruiting patients, what do you predict will be the results?


Another one bites the dust

It has been an article of faith, enshrined in the mnemonic MONA, to give supplemental oxygen to patients having heart attacks. (MONA is morphine, oxygen, nitrates, aspirin) The idea of course is that heart muscle is dying during a heart attack because a blood clot causes a lack of blood flow and oxygen deprivation.

A recent randomized controlled trial has shown that giving oxygen makes things worse!

AVOID trial

With this, we should drop the “O” in MONA . Maybe it should be “MAN.”

An issue of the sharp end several years back that I contributed to covered this very issue, along with other previous bits of dogma that have been overturned.

Room for thought: Is it possible that healing mechanisms after acute coronary syndrome have evolved to perform best in the absence of extra oxygen?

UNM School of Medicine Elective

Welcome UNM SOM students to the University of New Mexico Evolutionary Medicine Elective (CLNS 835).

Readings and assignments will be posted here weekly.

For the first session we will focus on the evolution of aging. As an example, I am more likely to admit a 70 year old with chest pain than a 20 year old. Increased risk of death and complications with age might be understandable as outcomes of evolved life history traits. Evolution informs nearly all aspects of human life history (which we will explain in the first session). Many diseases vary with age and senescence. We will explore these issues in light of this recently announced anti-aging $1 million prize:


We are going to cover evolutionary hypotheses of senescence during this week. These hypotheses include antagonistic pleiotropy, declining power of selection, and mutation accumulation. Read carefully:  Evolution of aging concepts.

For writing assignment and discussion: do you think the Palo Alto “hack-aging” prize will lead to a substantial increase in human longevity? Using evolutionary principles, defend your answer.

Katina’s Defense

Just a brief reminder of my dissertation defense on Friday morning.

“Discovery of Novel Major Histocompatibility Complex Class I Genes Unique To Marsupials and Monotremes”

Time: 9:00 am
Date: December 5th, 2014
Place: Castetter Hall, Room 107

Refreshments and snacks will be provided.

I hope to see you there!


Final Presentations

Start this Tuesday!

Here’s what to expect:

Be sure to click on the link that describes the grading rubric and has a template that you can (or not) use.

Here is what I will be looking for: Try to spend more than 50% of your time talking about the evolutionary implications or concepts surrounding your topic. Some previous students have started their presentations with a lot of detail about proximate causes and leave their discussion of an evolutionary hypothesis until the last slide or two. That is not good. While I am listening to your presentation, I am waiting for the evolutionary hypothesis and evidence that you understand how to consider diseases in terms of evolution and natural selection. So make sure that evolution appears early in your presentation.

Try to include at least one alternative hypothesis. Read the link for more details.

Finally, practice your presentation, preferably with a timer. You all did well during your Journal Club presentations, but I understand the temptation to talk longer than you anticipated when you are excited about a topic. We will be under a pretty strict time constraint. I want you to aim for 12 – 15 minutes per presentation.

Good luck


Pacifying our microbes – a way to prevent resistance evolution

On November 25th, Jon Femling MD PhD will give a lecture for the Evolutionary Medicine class. His talk will focus on the topic of microbial virulence evolution, pointing out several novel ways to prevent the evolution of antibiotic resistance using molecules that target virulence without killing microbes.

Jon Femling MD PhD

Dr. Femling will discuss ways to inhibit quorum sensing and virulence  gene expression in Staphylococcus aureus, transforming this often deadly bacterium from an enemy to a friend. Since this strategy does not kill the bacteria, it is expected to limit selection for antibiotic resistance.

Targeting Agr Quorum Sensing Inhibition in Staph


His recent paper can be read in its entirety here:

Selective Chemical Inhibition of Agr Quorum Sensing

Extra reading: Evolution of virulence. Ewald PW. 2004. Infect Dis Clin N Am (18)

There is no writing assignment this week. Work on your final presentations.