S(k)eptic.. Our resistance to activated Protein C

Introduction & Summary

Probably the news in sepsis is the outcome of the 'PROWESS' trial, which looked at the effectiveness of activated, recombinant human Protein C in the treatment of sepsis. Another name for this agent is drotrecogin alfa (activated). The trade name is Xigris. In the published results (which grabbed the front page of the 8 March 2001 issue of the prestigious New England Journal of Medicine), it is alleged that for the first time, a recombinant biological has altered mortality in sepsis. The potential of such an agent would seem immense, but we're not so sure that it should now be used widely, for the following reasons:

  1. It has recently come to light that the authors of the above mentioned article were somewhat 'economical with the truth'. Specifically they failed to reveal that (a) Trial entry criteria were changed halfway through the trial (b) The drug formulation was altered at the same point in the trial, allegedly without an affect on drug properties. Such changes only came to light during the process of FDA review.

  2. Following the trial's "mid-life crisis", the number of 'Do Not Resuscitate' orders documented on control patients far exceeded the number of such orders in patients receiving the trial drug. This could indicate that the patient profile in the two groups differed substantially, but of course could be related to the effectiveness of the drug, something which is not discussed in the NEJM article!

  3. As has been skilfully pointed out elsewhere, study implementation of other key therapies of sepsis (notably antibiotic therapy) may have been sub-optimal. (If you're in a rush, you can do no better than simply scan through Jeffrey Mann's incisive comments);

  4. Although an approximately six percent improvement in 28-day mortality was demonstrated, this does not appear to have been carried over into hospital discharges. It would appear that most of the 'extra survivors' were still languishing in hospital at 28 days, and no further follow-up is provided.

  5. The NEJM article states that: "Prospectively defined subgroup analyses were performed for a number of base-line characteristics, including the APACHE II score .. A consistent effect of treatment with drotrecogin alfa activated was observed .." In contrast, it was shown in the FDA review that certain patients did worse - notably those with lower APACHE II scores. One could argue that this is an artifact of subgroup analysis, but the FDA recommended that the drug be used in those "who have a high risk of death (e.g., as determined by APACHE II)". Conversely, the exclusion criteria for the trial (particularly the revised criteria instituted half-way through the trial) prevent us from knowing whether the drug is effective in many patients with particularly severe disease. We are left to use the drug only on those with (presumably) "middle of the road" sepsis!

  6. Although apparently a substantial clinical trial, the PROWESS trial is simply one trial. There have been numerous published clinical trials of biologicals in the management of sepsis. With usual cutoff "significance values" of p = 0.05, we might expect one trial to randomly come up "significant" just by virtue of the fact that so many trials have been performed, even if we ignore the possibility that yet more trials with negative outcomes might never have been published! We need at least one more trial, preferably an independent trial, that confirms the value of activated Protein C.

Note that in the above, we have not addressed the vexing question of cost. Although the drug will almost certainly be very pricey (Figures of US$ 6800 per course have been quoted), it is surely a matter of both individual and institutional decision as to whether the therapeutic value of the drug justifies the cost. In this regard, we note that the number of patients one has to treat in order to save one life is sixteen, so effectively the cost should be multiplied by sixteen - approximately 100 000 dollars per life saved (95% confidence intervals $70 000 to $325 000 per life saved). We also have no information on cost-effectiveness of the drug - as mentioned above, it would appear that patients whose 'lives were saved' by the drug did not necessarily achieve hospital discharge at 28 days.

We have also not invoked the haemorrhagic side effects of the drug as a reason not to use it, as the increased incidence of haemorrhage was fairly small (about 1.5%), and if the drug is truly effective in preventing death, such side effects may be a worthwhile cost to pay. Of concern in this regard is that, according to the FDA, the study incidence of haemorrhage may not reflect the true rate of intracranial haemorrhage.

On balance, it is our opinion that there are enough substantial, unanswered questions about the study to preclude our use of the drug in the management of patients with sepsis, regardless of the availability or price of the drug. Please note that this is a personal opinion. (You draw your own conclusions)! In the following sections, we look at some of the issues in more detail.

A brief review of Protein C

Protein C is really rather interesting. Protein C is present in the circulation in inactive form, and is activated (usually at the site of a clot) by a complex of thrombin and an endothelial protein called thrombomodulin. Unsurprisingly, we call the activated form of protein C, activated protein C, often abbreviated to APC. We have only recently begun to appreciate the important role of the protein C endothelial receptor {PCER} in Protein C activation - the PCER speeds the activation of protein C by thrombin+thrombomodulin about five times. [Blood 2001 Mar 15;97(6):1685-8]

When activated, protein C has, or is claimed to have, several effects. These include:

  1. Inhibition of clotting. APC has potent inhibitory effects on factors Va and VIIIa (and a lesser effect on the unactivated forms).
  2. Promotion of fibrinolysis. APC also inhibits something called PAI-1 (plasminogen activator inhibitor 1), and by so doing, indirectly promotes plasmin activity, and thus clot lysis.
  3. Anti-inflammatory effects. These appear complex.
  4. APC may also inhibit endothelial cell apoptosis.

Let's look at some of these effects in a little more detail. (Alternatively, you may wish to skip the technobabble, and find out about experimental sepsis). Claimed anti-inflammatory effects include:

  1. "inhibition of TNF production by monocytes";
  2. "blockade of leukocyte adhesion to selectins";
  3. "limitation of thrombin-induced inflammatory responses within the microvascular endothelium".

The anti-inflammatory properties of Protein C have recently been reviewed by Esmon in a supplement to Critical Care Medicine published under an unrestricted grant by Eli Lilly. If we in addition look on PubMed (using the search "Protein C[mh] AND monocytes[mh]") we find several articles:

One possible mechanism by which APC inhibits leukocyte adhesion to endothelium is fairly arcane. The theory is that endothelial cell protein C receptor {EPCR} binds proteinase 3, itself bound to leukocyte Mac-1 [Esmon]. In addition, it's been well shown that protein C inhibits E-selectin mediated cell adhesion, probably by virtue of its "polylactosamine" oligosaccharide residues. {We would be most interested in finding out whether the relevant residues are indeed present on recombinant APC, as we know that the residues on the synthetic agent differ from the naturally occurring protein C}. It's interesting to note that if, in an animal model of sepsis using intravenous injection of sublethal amounts of E. coli, we block binding of protein C (and activated protein C) to the EPCR, then the animals die [Blood 2000 Mar 1;95(5):1680-6]. In such models, the picture is of overwhelming sepsis.

Turning to PAI-1, if we search PubMed using "Plasminogen Activator Inhibitor 1[mh] AND Protein C[mh]", and weed out the irrelevant rubbish, we find one or two articles of note. Perhaps the most interesting article is a recent one from Rezaie [J Biol Chem 2001 May 11;276(19):15567-70] who says that "[The hypothesis that APC is profibrinolytic due to its inhibition of PAI-1], however, has not been well established, since the concentration of PAI-1 in plasma is low, and its reactivity with APC is very slow in a purified system." He shows that vitronectin greatly enhances the reaction of PAI-1 with APC.

In summary, there is some evidence that both protein C and activated protein C have "anti-inflammatory" effects, although the in vivo relevance of the demonstrated effects is open to question. It would seem that many of these effects require activated protein C to be present before the onset of the insult that causes cytokine production! Such a requirement would suggest that these effects are irrelevant to the clinical management of sepsis, but let's turn to experimental studies of APC in sepsis.

APC in Experimental Sepsis

The study that ignited interest in APC for sepsis is a very carefully performed and well-documented one from the Journal of Clinical Investigation [J Clin Invest 1987 79(3) 918-25]. In this study, conceived after it was noted that low dose thrombin infusion paradoxically ameliorates the effects of E. coli-induced sepsis, APC infused together with the E. coli unequivocally protected against death from sepsis. Even more convincingly, the authors demonstrated that use of mouse antibodies that blocked protein C activation resulted in worse outcomes, and even previously 'sublethal' doses of E. coli proved fatal.

I have only two concerns about the above study - firstly, that I can find no mention in the study of fluid resuscitation (if any) administered to the animals, and secondly, that the APC was given together with the E. coli infusion. In clinical medicine, we usually cannot anticipate the sepsis. If, for example, the anti-inflammatory effects of APC on monocytes depend on the APC being present on the cells when the inflammatory stimulus (e.g. endotoxin) hits the cell, then giving APC after the event may be too late. I am not aware of any animal studies of APC where the sepsis was induced, and then, later on, APC was infused, which would be similar to what happened in the PROWESS trial.

The Drug - A note on Xigris

A substantial PDF file is available for download from the website xigris.com. The document (which is also obtainable from the FDA website) describes the pharmacology of activated protein C, indications and contra-indications, and administration. Why not take a look at the company's documentation, tempered by the FDA?

The Trial - an overview

Readers are encouraged to consult the results of the PROWESS trial, published in the New England Journal of Medicine of 8 March 2001 Here's the abstract from PubMed:

Efficacy and safety of recombinant human activated protein C for severe sepsis.

Bernard GR, Vincent JL, Laterre PF, LaRosa SP, Dhainaut JF, Lopez-Rodriguez A, Steingrub JS, Garber GE, Helterbrand JD, Ely EW, Fisher CJ Jr; Recombinant human protein C Worldwide Evaluation in Severe Sepsis (PROWESS) study group.

Division of Allergy, Pulmonary and Critical Care Medicine, Vanderbilt University School of Medicine, Nashville, TN 37232, USA. gordon.bernard@mcmail.vanderbilt.edu

BACKGROUND: Drotrecogin alfa (activated), or recombinant human activated protein C, has antithrombotic, antiinflammatory, and profibrinolytic properties. In a previous study, drotrecogin alfa activated produced dose-dependent reductions in the levels of markers of coagulation and inflammation in patients with severe sepsis. In this phase 3 trial, we assessed whether treatment with drotrecogin alfa activated reduced the rate of death from any cause among patients with severe sepsis. METHODS: We conducted a randomized, double-blind, placebo-controlled, multicenter trial. Patients with systemic inflammation and organ failure due to acute infection were enrolled and assigned to receive an intravenous infusion of either placebo or drotrecogin alfa activated (24 microg per kilogram of body weight per hour) for a total duration of 96 hours. The prospectively defined primary end point was death from any cause and was assessed 28 days after the start of the infusion. Patients were monitored for adverse events; changes in vital signs, laboratory variables, and the results of microbiologic cultures; and the development of neutralizing antibodies against activated protein C. RESULTS: A total of 1690 randomized patients were treated (840 in the placebo group and 850 in the drotrecogin alfa activated group). The mortality rate was 30.8 percent in the placebo group and 24.7 percent in the drotrecogin alfa activated group. On the basis of the prospectively defined primary analysis, treatment with drotrecogin alfa activated was associated with a reduction in the relative risk of death of 19.4 percent (95 percent confidence interval, 6.6 to 30.5) and an absolute reduction in the risk of death of 6.1 percent (P=0.005). The incidence of serious bleeding was higher in the drotrecogin alfa activated group than in the placebo group (3.5 percent vs. 2.0 percent, P=0.06). CONCLUSIONS: Treatment with drotrecogin alfa activated significantly reduces mortality in patients with severe sepsis and may be associated with an increased risk of bleeding.

Probing Deeper

We are indebted to Jeffrey Mann, who put a searching and well thought out analysis of the PROWESS trial onto the Emergency Medical Practitioners List Server. You are encouraged to read both his analysis, and the FDA documents on which it is based - both the transcript of the proceedings, and the FDA briefing document. Here are a few relevant quotes from the transcript, but you are encouraged to read it in its entirety (all ~ 300 pages) - as you'll see from the following quotes, this is no mean task! All quotes are in italics. "BDS2+" is the altered drug formulation. (If however, you want to skip to our conclusion, then click here)!

    Amendments to the protocol

    These are extensively discussed. Initially the alterations are underplayed, but then it emerges that they were indeed substantial. Here are just a few quotes, in which matters such as the 28 day survival and "DNR" orders also come up. The quotes do not follow one another directly in the FDA document:

  1. As I mentioned there was one amendment to the protocol. The amendment occurred very early in the course of the study
  2. DR. MACIAS: I think there are two points to it. One is as a point of clarity, we didn't exclude patients with underlying non-sepsis-related diseases. What we excluded or who we excluded were patients with non-sepsis-related diseases in whom the investigator would assess the patient as being at high likelihood of dying within the 28-day study period. That was what we excluded.
    DR. FLEMING: Is your label accordingly excluding such patients? I don't see that it is.
  3. DR. FLEMING: At least I would like to call to the FDA's attention two issues that to me are major concerns. One is that it is apparent that Lilly was assuming the responsibility as opposed to an external committee for early safety monitoring and secondly, there was access to this code within Lilly, and we had a change in the primary end point during, well into the course of the study.
  4. DR. MURRAY: Sort of two questions. One is a follow-up on the previous question relating to the proposed labeling by Lilly. Would the proposal be that people with organ failure more than 48 hours not be administered the drug since that is how it was studied and then the second question which is totally unrelated but is more for curiosity, if you have any longer-term follow-up to the 28 days to see if people that did survive that had not gone back home were back to a more functional status?
    DR. MACIAS: To answer the first question, from a restriction standpoint we would not propose that we would restrict it 48 hours, but the label would reflect the fact that patients were enrolled in the study within 48 hours of the onset of their organ dysfunction. The reason for that is I don't think we believe that if a patient has organ dysfunction for 47 hours he would be eligible and for 50 hours he would be ineligible which is I think very different than TPA and stroke where if you have a stroke syndrome for 3 hours you might benefit and when you have stroke syndrome for 6 hours you don't benefit. So, I think there would be a reflection, we would recommend reflection of the duration of organ failure in the clinical trial section of the label but not a restriction for the reasons I have expressed. With respect to follow-up beyond 28 days, we have currently initiated the follow-up protocol to EVAD. So, it will follow up all survivors in the EVAD protocol up through I think for some patients a minimum of 1 year but for patients enrolled early follow-up will be longer.
  5. DR. SUFFREDINI: I guess there are at least four factors that changed between Part A and Part B of the study and in terms of the change in the DNR rate in the treatment group that diminished significantly but that wasn't shared by the placebo group which certainly may have a significant effect in terms of the 28-day mortality. Can you comment on that?
    DR. MACIAS: Yes, I think it is very difficult to interpret the DNR rates between the two groups because the DNR rate may reflect the fact that you gave them placebo and not effective therapy. The incident rate of making someone -- DNR may go down in the setting of an effective therapy.
  6. DR. O'FALLON: The coincidence of the change in the protocol at approximately the same time that the first interim analysis was performed has bothered me as I read this material, and the fundamental question is, and I assume that the Committee when they did the second interim analysis did understand that there had been major changes in the protocol and attempted to take that into account as they made their recommendations. Is that correct?
    DR. MACIAS: Dr. Opal, would you like to address that question, please? Dr. Opal was the Chairman of the DMSB.
    DR. OPAL: Yes, we were aware of the protocol amendment and did take that into consideration in looking at the data.
  7. Finally, before I conclude with the summary of efficacy I would like to focus on the changes that were made to the protocol by the sponsor. These occurred while the study was still blinded and before the first interim analysis was conducted. The sponsor made two sets of changes in July 1999. They altered their analytic time and they clarified their inclusion and exclusion criteria and they eliminated protein C deficiency and septic shock from the Cochran-Mantel-Haenszel test. The sponsor changed the inclusion and exclusion criteria to clarify certain parameters. This was done to clarify definitions and better eliminate patients with chronic or comorbid disease. These patients were not likely to respond from acute therapy of severe sepsis in the sponsor's evaluation. Listed here are some of the criteria that were modified. These include excluding patients that were more likely to bleed such as patients with esophageal varices or cirrhosis. They also excluded patients that were more likely to die and patients with other underlying disease such as malignancies and, also, they further clarified their organ failure eligibility for the trial. These are the results of the changes between the original and amended protocol. There were a few relative modest differences between the original as compared to the amended version of the protocol. In the amendment there were fewer patients with malignancies that were immunosuppresed that had the withdrawal of life support, with chronic APACHE health points and with non-sepsis-related disease as well as at nursing home facilities.
  8. We, also, looked at the differences between the number of "do not resuscitate" orders under the original versus the amended version of the protocol. As can be seen the placebo rates were similar. However, the number of patients in the first half of the study or in the original protocol on rhAPC it was 16 percent compared to 10 percent in the second half of the study or under the amendment. This may reflect differences in mortality as on the next slide. Here as previous touched upon this table displays mortality data stratified by the original and the amended versions of the protocol. As we can see there were 720 patients that were enrolled in the original protocol and 920 under the amendment. The placebo rates between the two protocol versions are similar. However, in the rhAPC arm there was a 28 percent difference in mortality under the original protocol compared to 22 percent in the amended version of the protocol. As noted we, also, have the P values here of .0055 or .057. Under the original protocol and under the amended protocol we have a P value of .00012. It is, however, important to review not just the individual landmark analyses of the data for specific patients but to look over the entirety of the survival curves as a function of study day. Shown in this table as the sponsor has already shown are the mortality rates in each arm on day 1 and really on a day-to-day basis as the study was conducted. On the X axis we have the dates throughout the study. On the Y we have the mortality rates depicted for each study arm reflecting 28-day observation and therapy. Let me go ahead and point out a few salient features. Line A represents the time the first patient was enrolled under the amended version of the protocol. Line B occurs when the first interim analysis occurred and Line C when the second interim analysis occurred. Of note, there was a separation of the curves before the protocol amendment was implemented in favor of benefit for rhAPC. These curves continued to separate throughout as the study progressed. This curve was conducted in 1690 patients and slightly differs from the sponsor's curve presented earlier today as the numbers differ. The sponsor apparently was using a number of sites, of 99 sites and 1493 patients, and furthermore we conducted a sensitivity analysis evaluation for those patients that were enrolled in the original protocol, but would have been excluded from the amended protocol. There were a total of 81 patients or 11 percent that did not meet new inclusion criteria for the amended protocol that would have met the original protocol. Despite the 81 patients who did not meet the inclusion criteria they continued to show treatment effect. So, in summary there was a benefit in the 28-day all-cause mortality. The mortality rate on rhAPC was 24.7 percent versus the 30.8 percent on placebo with a P value of .005. Finally treatment benefit was more predominant in the following groups of patients, in the third and fourth APACHE quartiles compared to the first and second, in patients with laboratory evidence of DIC compared to those without and also in patients receiving heparin, oh, not on heparin, excuse me, compared to those receiving low-dose heparin, in age groups over 50 years of age compared to not and in greater than two organ failures at baseline compared to patients without or single organ failure and finally, in patients with shock compared to patients without shock.
  9. DR. MACIAS: Excuse me for one second. The BDS2+ material did not go through the same preclinical animal toxicology studies that the original material did.
  10. Antibiotic Therapy

    As Dr Mann points out, in the management of sepsis it may be critically important to get the antibiotic therapy right, and get it right quickly. Contrast these sentiments with:

  11. DR. LEGGETT: A couple of questions. In terms of the survival benefit, it appeared that most of the time in the Kaplan-Myer [Meier] curve that the survival benefit happened after the infusion. So, I would take that sort of physiologically to mean as you, I think tried to show that it was less organ dysfunction down the road, and you made the statement in, I cannot remember the slide where the post-5-day-infection rates, post-baseline infection rates were the same. Was the post-baseline mortality from those infections the same as part of that first question? The second part of that first question is when people went back and looked at these patients in terms of other routine care, were the assessments of optimal or appropriate antibiotics in the two groups controlled for or looked at to see if that could be a potential confounding variable much like the steroid question?
    DR. MACIAS: The answer to the very first question which is what were the mortality rates between treatment groups for patients having a post-baseline new infection, did we look at that, Jeff, do you remember? I don't think we looked at that. We can look at that for you today. To answer your second question, and that is the appropriateness of antibiotics, all of the case report forms were reviewed by our Clinical Evaluation Committee and appropriateness of antibiotic therapy was adjudicated by the CEC. Approximately 88 percent of patients received appropriate antibiotic therapy within 24 hours of the diagnosis of severe sepsis, and the proportion was equal between the two treatment groups, and by 48 hours 92 to 93 percent of patients had received appropriate antibiotic therapy.
  12. APACHE score and outcome

    APACHE score correlated with outcome - treatment benefits were seen mainly in those with higher scores:

  13. The mortality difference in percentage here presented in the third column from the right we can see there is virtually no difference in the mortality between rhAPC and placebo in the combined first and second quartiles; however, among patients in the third and fourth quartiles the combined APACHE difference or mortality difference, excuse me, was 13 percent.
  14. These patients that are in the lowest APACHE II grouping have about a 12 percent mortality. Even if in truth there was a 20 percent reduction, that translates into two or three deaths per hundred people versus the six that we have been told about for the broader group. Two or three deaths is certainly relevant, but it is a smaller overall benefit. That is presuming that there is, in fact, as the sponsor would argue or as somebody who says there is a homogeneous effect would argue, that assumes that we have a 20 percent reduction. Of course, as the data show, there is a 25 percent increase. So, there is a leap of faith there to argue that there would even be two deaths per 100 people. So, there is a smaller unmet need. It is still a significant need, but it is a smaller unmet need in that group. Secondly, there is some consistency of evidence that in these lower risk patients, the effect may be much less. Thirdly, there is a non-trivial safety risk. That safety risk is in part represented by these SAE bleeds. There were nine of them. Four of those people died. That is a 2 percent death rate. It may not be that those deaths with major bleeds were entirely mediated through the effect of the treatment, but if, in fact, they were substantially mediated so that you had four deaths, that is 2 percent, that would exactly offset the 2 percent benefit, which, in fact, we didn't see, but in a best case scenario, if we argued there wasn't an interaction, we would say would be there. Finally, and I don't have a good sense about this, but the additional evidence of the 1 1/2 percent risk of intracranial bleeds that was referred to by the FDA in the more recent data also becomes more of a concern in a low risk population. The lower the risk the population, the lower the overall net benefit. Hence, the bigger concern with the level of side effect.
  15. Morbidity at 28 days

    Many of the "extra patients who survived" were still in hospital at 28 days, with no indication as to their status:

  16. I would now like to switch gears and focus not only on mortality but on morbidity. Since 20-day all-cause mortality does not reflect all outcomes of treatment benefit it is important not only to look at the number of patients alive but at their treatment status and evidence of comorbidity. Shown here in this slide is a side-by-side comparison of morbidity and functional status at day 28 compared between rhAPC and placebo. Mortality is shown in red. In green we have a side-by-side comparison of ICU status at day 28. In yellow patients that are still hospitalized at the end of the study are shown. Violet is patients that are in a nursing home and navy in a discharged home. In the rhAPC arm there was as you can see approximately a 6 percent difference of fewer patients who died on study. At day 28 this group had about a 2-1/2 percent number of more patients that were alive in the ICU and alive in hospital. Also, about a 1 percent difference occurred in patients that were discharged to a nursing home or to home.
  17. What we can see is, and what we focused on is the difference in the red, that there are 6.1 percent fewer deaths which numerically is a difference of 49 additional deaths in placebo and yet the green and the yellow groups there are those that are hospitalized and there are exactly 46 additional treatment patients in the hospital. So, essentially what this translates into then is a 6.1 percent reduction in death, but of that 6.1 percent 5.1 percent are in the hospital. So, it is not as though we are in essence increasing by 6.1 the percent of people who are home. Essentially for the most part the prevented deaths are people who were in the hospital. Clearly it is better to be in the hospital than to be dead, and yet it is unclear exactly what this will translate into in terms of overall effect, and the reason this is important as well, and we will come back to this later on is that this study does meet the standard for strength of evidence for a positive study for a single positive study and that is of course, though on survival, and if you in essence though subtract off in some subjective sense the fact that you are hospitalized you could readily if you subtract off anything meaningful be moving to a level where this wouldn't be significant. Clearly the study isn't remotely significant in the number of people who are alive out of the hospital because there is only a 1 percent difference.
  18. Risks of bleeding

  19. There is a clearly identified increased risk of bleeding in patients treated with rhAPC. Intracranial hemorrhages were identified in only two patients treated with rhAPC in the Phase III trial. New data suggest this may under-represent the actual rate. Additional intracranial hemorrhages may go undetected in situations where CT scan for practical reasons is not performed. Thus, it is unclear what the true rate of intracranial hemorrhages may be. Other major bleeding events occurring in contained non-visible sites could be difficult to detect for the same reasons. Though major bleeding events were identified, the risk of these events remains somewhat uncertain. Thank you.
  20. Definitions of DIC

    There seems to be some controversy in the document about whether substantial evidence of "DIC" was present, or not:

  21. The majority of patients in this trial, over 90 percent had laboratory evidence of DIC at study entry. This was defined by the presence of at least two of four laboratory findings as the sponsor has previously defined today. There were 115 patients in which DIC was unknown or absent. In 113 patients of this group this was due to insufficient laboratory data available to determine DIC. Of the small number of patients there appears to be a limited treatment effect in this group.
  22. DR. SUFFREDINI: I wonder if Dr. Forsyth could address the issue. I guess I am concerned about your conclusion that the treatment was efficacious with laboratory evidence of DIC, and my concern is that the definition as used by the sponsor is certainly not robust in terms of the two out of four parameters that are used are really very common abnormalities that one would see in any patient that goes into ICU and certainly don't meet the criteria that most textbooks would use for DIC and what a practitioner would consider severe disseminated intravascular coagulation. So, it is not surprising since there is not a discriminating factor DIC the definition used was so broad. It is really not DIC. It is an abnormality in coagulation perhaps, but that is really not DIC, and so, my concern is in terms of labeling or in terms of how the practitioner would look at this and use it in someone with profound DIC. I am not sure if the data exist to tell us that that would be useful or not.
    DR. FORSYTH: I totally agree with you and as far as the no standard definition of DIC we certainly understand that and this is the way that the sponsor has proposed to address the issue of DIC but I agree with you on that. That is an issue. As far as labeling is concerned, we have been struggling with this issue


"It was on September 11th that this panel, I believe, was originally going to convene and, of course, was cancelled, as have many of the other plans of many Americans been cancelled from that date. But, hopefully, one thing that we can do here today is we can move forward with at least one weapon that we have to fight a serious problem in this country"

[Dr Smirniotopoulos, a pulmonologist and critical care physician,
speaking at the FDA Open Public Hearing on Xigris].

Dr Smirniotopoulos would appear to be encouraging us to start a "war on sepsis". If one is to fight a war, it's generally a good idea to know one's enemy, their strengths and weaknesses, and the consequences of their defeat. Unfortunately, in combatting sepsis, we are sure of none of these requirements. Although there are three quarters of a million new cases of sepsis every year in the USA, it is far from clear how many of the two hundred thousand who succumb have advanced diseases where aggressive therapy is inappropriate. If a new therapy for sepsis merely prolongs the misery of such patients, so that they languish in hospital instead of dying, we may have replaced the old 'Taliban' regime with something different, but no better!

There is great difficulty in accurately characterising sepsis. Bone's consensus criteria defined sepsis as "the simultaneous occurrence of the Systemic Inflammatory Response Syndrome and a confirmed infectious process". The heterogeneous mix of ICU cases who display such features is light years distant from laboratory baboons infused with E. coli. To make things even less certain, the widely-accepted consensus definition was modified for the purposes of the PROWESS study.

We are not certain why "septic" patients die, although microvascular dysfunction, coagulation abnormalities, relative hypovolaemia, myocardial depression and other major organ dysfunction probably all contribute. It is the author's personal belief that many patients who die from sepsis do so because the underlying source of infection has not been timeously and thoroughly addressed, resuscutation has been inadequate, or the patient has been rendered helpless by poor nutrition or advanced underlying disease. It seems excessively naive to believe that there is one "magic bullet" that will somehow miraculously solve such complex problems.

At the start of this webpage, we listed reasons why we felt that activated protein C should not yet be used widely in the management of sepsis. In the sections that followed, we supported this contention - there are several worrying questions about the study methodology and outcomes. In addition, the agent is not being used for replacement of a deficiency state - rather, it is being used pharmacologically. The study results indicate that APC is effective whether there is or is not evidence of an underlying coagulopathy, and regardless of Protein C levels. Such evidence (if it is to be believed) suggests that the 'mechanism of action' is related not to amelioration of coagulation defects, but to other effects, perhaps "anti-inflammatory" actions of the drug. The latter have however not been well-characterised, and (based on in vitro evidence) appear to work mainly early on in the activation of monocytes, a mode of action that cannot necessarily be extrapolated to the clinical state. Although such ignorance should not prevent us from using the drug if it truly works, a firm biological basis and robust pre-clinical work (both of which seem absent) would make me less skeptical about the drug. In addition, the therapy is not without risk, and the PROWESS study may have underestimated the incidence of intracranial haemorrhage.

Perhaps, after all, there are some similarities between sepsis and the events surrounding September 11th, 2001, although not in the sense that Dr Smirniotopoulos intended. We are uncertain about our 'enemy', and the 'weapons' we have at our disposal may, if used inconsiderately, result in damage to unwanted targets (remote villages and Red Cross depots, to continue the analogy), and even long-term adverse outcomes.

The real worry is that if the researchers involved in the PROWESS trial provide more information (for example, on the status of patients at 28 days in the control and Xigris groups), even if the results are published in, say, the NEJM, how can we believe them? What might have been left out? Perhaps Jeffrey Mann is correct when he suggests that all trials be preceded by a website open to public criticism! And maybe we should not trust any trial unless all data are open to public scrutiny, not just the data that the drug company saw fit to provide!


  1. Bernard GR, Vincent JL, Laterre PF, et al. N Engl J Med 2001 344(10) 699-709. Efficacy and safety of recombinant human activated protein C for severe sepsis.

  2. Esmon CT. Crit Care Med 2001 29(7S) S48-52. Protein C anticoagulant pathway and its role in controlling microvascular thrombosis and inflammation.

  3. Mann, Jeffrey. http://www.ucsf.edu/its/listserv/emed-l/4879.html
    Listserve For Emergency Medical Practitioners: EMED-L.

  4. Mann, Jeffrey. http://emguidemaps.homestead.com/files/xigrisebm.html
    The Xigris Affair: Searching for a new paradigm in the world of evidence-based medicine.

  5. FDA PDF licence for drotrecogin alfa. http://www.fda.gov/cber/label/droteli112101LB.pdf

  6. FDA transcript. http://www.fda.gov/ohrms/dockets/ac/01/transcripts/3797t1.doc

  7. FDA briefing document. http://www.fda.gov/ohrms/dockets/ac/01/briefing/3797b1.htm"