DIC
July 12, 2003
14:00 to 18:00
Hall 5
The International Convention Center, Birmingham
Chairman: K. Hoots, USA
Co-chairs: I. Bokarew, Russia;; M. Levi, The Netherlands; J. Nielsen,
Denmark;
N. Sakuragawa, Japan; C-H. Toh, UK; H. Wada, Japan
The following report and addendum summarize both an interim meeting of the
subcommittee co-chairs held in Houston in February 2003 and the subcommittee
meeting held in Birmingham in July 2003. The latter meeting summarized
implementation and recommendations concerning the DIC algorithm and committee
progress to the entire membership of the DIC subcommittee and was verbally
endorsed by those present.
This is a summary of the DIC Subcommittee of the Scientific Standardizations
Committee of the International Society of Thrombosis and Hemostasis held
in Houston, Texas, February 14-16, 2003 at the University of Texas Houston
Health Science Center.
The agenda which was circulated to the committee prior to the meeting consisted
of four (4) major areas of discussion. The first was a review of the
DIC Subcommittee Algorithm which was published in the Journal Thrombosis
and Haemostasis in November 2001. The intent was to examine sensitivities
and specifities of the algorithm based on pilot studies done in at least
3 sites. Also discussion was to center on how and if changes needed
to be undertaken to the algorithm. Second there was review of updates
from the literature and experience with regards to clinical trials and potential
trials. In particular the PROWESS trial and the KYBERSEPT trial were
discussed in great detail. Number 3 on the agenda was any new endeavors
to be undertaken by the committee with regard to the algorithm or other new
efforts on behalf of the subcommittee. The final intent was to set
the agenda for the Birmingham meeting of the DIC Subcommittee at ISTH in
July 2003. Attendees at the meeting were the following: Dr. Keith Hoots
was the host, excused was Dr. Marcel Levi who because of commitments in his
home country of the Netherlands was unable to attend, Dr. Cheng Hock Toh
from Liverpool, Dr. Hideo Wada from Tokyo, Dr. Carl Erich Dempfle from Hamburg,
Dr. Fletcher Taylor from Oklahoma City, and Dr. Mathias J‡rs from Aventis
Behring (meeting sponsor) were the attendees present.
Introduction: The ultimate aim of this Subcommittee by consensus is
to improve outcomes in individual patients who have diseases that lead to
Disseminated Intravascular Coagulation (DIC). It was agreed that overt
DIC is the final common pathway in that in many if not all cases it is initiated
or at least exasperated by the enzyme thrombin which behaves in a very chemically
promiscuous way. The consensus approach to be taken from the meeting
was as follows:
1) to define the process to apply the DIC algorithm across wide spectrums
of clinical diseases and geographies and multiple medical specialties.
Number two was to validate the scoring of the algorithm in diverse groups
with both overt and non-overt DIC. Initially next steps about the algorithm
were discussed in significant detail. The first effort was to examine
validation progress on overt DIC. Data was presented from Liverpool,
Tokyo and from Hamburg related to efforts to apply the algorithm in clinical
populations with DIC. This was an update of information that was presented
at the last DIC Subcommittee in Boston in August 2002. In addition,
some new information was provided by Dr. Wada with regards to comparing the
DIC score of our committee with an ICU/ER score which was an outgrowth of
the original Japanese Ministry of Health DIC score which served as one of
the basis for our developments of this algorithm.
2) We discussed the status of possible multicenter validation of the overt
DIC algorithm and felt that we would benefit greatly from the presence of
Dr. Levi to discuss protocol implementation practices and to reach among
ourselves definitive statements concerning inclusion, exclusion and analysis
criteria to be utilized for such studies.
3) We wish to define the "message" to be conveyed to the broad medical community
concerning this ISTH algorithm and to continue updating comparisons between
it and the ICU/ER algorithm in Japan versus the traditional Japanese Ministry
of Health and Welfare algorithm. Dr. Wada will take the lead on this
and will continue to make such comparison in groups of multiple types of
patients with multiple insults and injury.
4) With Dr. Dempfle as the discussion leader we examined the role of fibrin
markers in the scoring system itself. It was recommended that cutoffs
between 0, 1, and 2 that score on the DIC overt and non-overt scale adopt
ranges that are compatible with the SSC Fibrinolysis Subcommittee on which
Dr. Dempfle sits. Specifically for D-dimer we proposed that if it was
1 microgram per mil > than 1 it would score 2 but if it was £ 4;
greater than 4 micrograms per mil we would score 3 in the overt/non-overt
scale was the recommendation that was entertained. Further discussion
should follow.
5) We also proposed d-Dimer as the ideal fibrin marker. We left open
the option for fibrin degradation products or new fibrinolytic or fibrin
markers to be employable in the algorithm itself. From the data provided
by Dr. Dempfle we recognized the need that a specific marker of excess fibrin
degradation be identified and efforts to validate it for possible implementation
be undertaken. This is critical for future studies to determine whether
proteolysis of fibrin confers added predictability to what is already conferred
by the presence of markers of fibrin deposition.
6) We queried whether an additional list of fibrinolytic markers would be
beneficial in the algorithm. No consensus was reached at this time
we will plan to return to this discussion.
7) It was noted that there were a number of confounding phenomenon making
utilization of fibrin markers and fibrinolytic markers more problematic:
for example, increase plasminogen activator inhibitor 1 (PAI-1) correlates
with DIC at least in sepsis and would be expected to actually decrease fibrin
degradation products and therefore may compound the interpretation in the
acute setting. Further examination of this possible confounding effect
needs to be undertaken with the consensus of the Subcommittee.
8) We entertained the question "Does the scoring for Fibrin Marker(s)
need to change in light of heterogenecity of assay results of D-dimer?"
Specifically since multiple assays exist for D-dimer it is important that
validation for each specific marker be present at the site which proposes
to use that D-dimer test. The other thing about overt DIC was that
we agreed that it could not be used as a trend marker unlike that the situation
that makes it with the non-overt DIC.
The next part of the committee presentation centered on clinical trials in
DIC that have been completed. Both the PROWESS trial and the KYBERSEPT
trial were reviewed on a broad basis. It was noted of particular import
to this Subcommittee that both of these trials had sepsis as the inclusion
criterion rather than DIC, even though the results are most frequently identified
as either success or failure to treat DIC. It was the consensus of
the group that future trials should in fact have DIC as an inclusion criterion
if in fact the conclusions were going to be drawn concerning success or failure
of the treatment of DIC.
With regards to the role of DIC algorithm in future clinical trial designs
an extensive discussion was undertaken. It was felt that several things
needed to be improved in the algorithm scoring system in order to maximize
its potential for such an application in trial design. Specifically
two issues were noted.
1) It was felt that markers of neutrophil activation such as elastese as
a specific indicator of cell interaction and upregulation in response to
injury would be needed to meet the needs for certain trials.
2) In addition, other clinical parameters need to be examined carefully.
It is essential that critical ascertainment for an individual patient must
be made about where along the non-overt/overt continuum he/she is at the
time of initiation or their entry into a trial. It was felt that this
was a critical point in order to assess whether in fact progression/regression
or a static situation had been achieved in response to therapy. The
"TIME" dimension for progression along the continuum is absolutely critical
for this assessment. That is to say, we need to be able to provide
measurements as prescribed by the Algorithm in a sequence of clinical/laboratory
events. This is so that the definition for inclusion in the trials
can identify targeted patients who fall within a narrow window along the
continuum or at least so that they can be stratified during the analysis
accordingly.
Other issues included the following: the biokinetics need to be defined better
for categorical microangiopathic states. Efforts of Dr. Wada and his
other Japanese colleagues have looked at static determinates of DIC and microangiopathic
states, comparing for example HUSTTP with full-blown DIC with other inflammatory
states that are not microangiopathic. This provides a very useful scenario
for validating our algorithm. It is important however that the sequence
of either progression or lack of progression in molecular markers be defined
for each of these generic clinical situations so that in clinical trials
the right patients can be included or excluded appropriately.
2) We need to design a way so that the impact of therapy on the biokinetics
of the DIC continuum can be assessed. This will of necessity require
further elucidation of the transition between overt and non-overt DIC.
3) It was felt that future directions would require assessment of genomic
differences perhaps proteomic differences in individual subpopulation
groups and the potential impact that each may have on propensity to progress,
regress or remain static in response to therapy. One of the things that clearly
is needed (at least for each of the markers to be employed) is a consistent
longitudinal graph of the biological marker parameter versus time with an
attached confidence interval. This needs to be performed longitudinally
for each of clinical conditions so that investigators can best decide where
in the continuum these research subjects lie.
Finally, in order to expedite this and other uses with regards to both the
non-overt and the overt algorithm, specific recommendations to modify and
improve the Algorithm were made at this meeting of the Subcommittee.
These consisted of three specific recommendations:
1) To specify specifically which fibrin marker assay should be used for example
D-dimer;
2) DO NOT mix fibrin marker assays in the same study unless that is
the comparison for the study design to be examined; and
3) change the prothrombin measurement to the Prothrombin ratio as defined
by QUICK this is in distinction to the INR. It is calculated at 70%
of the normal range for the specific laboratory control of the laboratory
used in the assessment.
The final endeavor for this committee meeting was to outline the next steps
for further assessing and then implementing the algorithm. With regards
to non-overt DIC, since DIC has been summarized above, the following issues
or recommendations were made:
1) As a major criteria, change "soluble fibrin/fibrin degradation product"
in the algorithm to the term "Fibrin Marker" and see the discussion above
under overt.
2) A cut-off for the D-dimer needs to be defined in terms of milligrams per
ml in most assays twice the upper limits of normal range should be used as
a defining principle.
3a) For the prothrombin time use the QUICK ratio as defined above in the
overt discussion.
3b) Use only platelet counts that are pre-transfusion because obviously there
is a confounding effort in the analysis if post transfusion platelet counts
are used.
4) If anti-thrombin concentrates or activated Protein C concentrates
or non-activated Protein C concentrates have been utilized to treat the patient
one cannot use the specific criteria for entry into this algorithm.
5) In recalculating the non-overt score one needs to take into consideration
specific criteria about whether and what impact the use of fresh frozen plasma
may have had in the starting score and in the progression of the score.
One of the questions that was raised was whether such assessment should be
part of our ongoing trend analysis in the implementation of the algorithm.
6) There needs to be a vigorous attempt at "normalizing" the biological
marker data parameters. For instance, in the non-overt we indicate
that either Protein C or Antithrombin can be used to score over time.
It was felt by the Subcommittee that we need to be consistent that one cannot
measure one parameter on one point in time and then switch to another parameter
on another point in time and score. It must be decided apriori which
biological markers are to be implemented for the full duration of the use
in the individual patient in the non-overt algorithm.
7) It was felt that thrombin Antithrombin or TAT should not be utilized because
it cannot be done in real time. It can only be done after the fact.
8) It was felt that Wave Form Analysis could be used prospectively
but that it should be normalized as an extra test in the scoring so that
both validity could be felt and so that over sampling of biological parameters
do not increase the statistical likelihood of error.
9) It was felt strongly that for the non-overt algorithm we need an
endothelial and a polymorph nuclear neutrophil activation set of markers.
These need to be able to be done in real time. Examples that might
be applicable include thrombomodulin and elastese; markers of fibrinolytic
inhibitors such as PAI-1 the latter would be very useful for use in the non-overt
algorithm in particular because of the demonstrable role that it seems to
play in the evolution of sepsis in particular.
10) We need to be much more aggressive in defining outcome measures in our
effort to pilot the non-overt algorithm. We need to know
the answers to the following specific questions: 1) How many of those who
enter into the non-overt algorithm proceed to overt DIC by the criteria we
have defined? 2) For both overt and non-overt uses of the algorithm
we need to know exactly how many people die and whether they die from DIC
or from some other cause in order to validate the overall use of the algorithms.
3) We need to be able to show how many individual subjects show significant
improvement as they go from the use of the non-overt algorithm towards the
use of the overt algorithm. For example, if they start with DIC what
is the impact of therapy; if they start without DIC what is the impact of
therapy in preventing or allowing progression to DIC. We clearly need
prospective data to confirm these retrospective analyses that have been done
to date. And finally
11) we need to know the timing of the use of the algorithm specifically,
we need to know (1) how often can you reapply the algorithm if at all; (2)
how many times can you use it within a protracted course within the same
individual; and (3) over what period of time over what duration of time can
the algorithm be applied? The consensus was that the first step the
algorithm should not be applied more often than daily because of the problems
created by repeating sampling error that this would introduce into assessing
the validity of the algorithm.
The final undertaking of the committee was to recommend a program outline
for our subcommittee meeting of ISTH in Birmingham. Some discussion
was undertaken. Since we hadn’t previously discussed it it was
felt that it might be very helpful to look at both protease activated receptors
those of particular interest with thrombin activation such as PAR-1and PAR-2.
Drs. Hoots and Taylor recommended Shaun Coughlin from the University of California-San
Francisco who has written an excellent review on this in Nature. Dr.
Hoots agreed to contact Dr. Coughlin with regards to his possible availability
for speaking at our meeting. Two other suggestions were made, one was
Dr. Ruf from the Scripps Institute in La Jolla who is also an expert in the
PARs and Dr. Nigel Mackman also from Scripps who is an expert in phospholipid
services where proteases bind and the importance of enzymes such as "scrambalase"
that change the phospholipid services in response to injury signals.
One of the things that we thought would be interesting is reviewing recent
data that the endothelial Protein C receptor signals through a PAR-2 receptor
to examine how this might somehow be related.
Our final recommendation is that we establish a World Working Group for DIC.
It would be expected that the committee co-chairs from our committee would
work with colleagues in the field of critical care, sepsis and trauma and
other related fields for which the algorithm would have particular relevance.
Dr. Dempfle and Dr. J‡rs recommended the German Sepsis Group and Dr. Taylor
recommended Dr. Gary Kinasewitz, MD at the University of Oklahoma who is
a Pulmonologist and an Intensivist who has worked extensively with him as
potential members of such a working group. These issues decided, we
adjourned the meeting and the intent was for Dr. Hoots to summarize the proceedings
and distribute them to the subcommittee co-chairs first and after the refinement
perhaps to circulate this information at our meeting in Birmingham.
Addendum
The meeting of the subcommittee was held on Saturday, July 12, 2003 14:00
– 18:00 at Hall 5 of the ICC in Birmingham, UK. Dr. Matthias Reiwald
presented a mini-symposium lecture entitled "Interaction of Coagulation Proteases
with cellular receptors and phospholipid cell surfaces: Proteases Activated
Receptors (PARs) and Inflammation, Possible dysregulation in DIC."
Dr. Toh presented the Liverpool experience using the overt and non-overt
algorithms and Dr. Wada presented comparison studies between the SCC algorithms
and the ongoing algorithms undergoing modification from the Japanese Scientific
Coagulation Committee/Japanese Ministry of Health and Welfare.
In addition Dr. Marcel Levi applied the overt algorithm to the data from
the PROWESS study which indicated that the overt algorithm was very predictive
of mortality in the APC treated cohort. In addition Dr. Carl-Erik Dempfle
of Mannheim, a liaison from the Fibrinolysis Subcommittee presented a simple
data on the role of marker on fibrin cleavage in the DIC algorithm and presented
the first quantitative measures included recommendations, of how d-Dimer
levels should be stratified for scoring in the overt DIC algorithm.
The SCC attendees then discussed optimal plans for future studies and modifications
to the SCC approved algorithms. Plans for a future meeting prior to
or before SCC meeting in Venice were discussed by the DIC SCC subcommittee
chairs and will be coordinated by the chair.
Respectfully submitted: W. Keith Hoots, MD