Cape Town 2008

3 March 2008

Open Letter To:

Tony Absalom, Luc Barvais, Valérie Billard, Alain Borgeat, Thomas Bouillon, Sorin Brull, Nassib Chamoun, John Coetzee, Ignacio Cortinez, David Crankshaw, Albert Dahan, Peter Davis, Tom de Smet, Jay Deshpande, Phil Devlin, Anthony Doufas, Talmage Egan, Frank Engbers, Pedro Gambús, TJ Gan, Tony Gin, Peter Glass, Iain Glen, Harald Ihmsen, Ken Johnson, Tomiei Kazama, Gavin Kenny, Kai Kuck, Frank McGowan, Claude Meistelman, Charles Minto, Erik Olofsen, Quirino Piacevoli, Donald Prough, Johan Raeder, Douglas Russell, Stephane Ruton, Thomas Schnider, Peter Schumacher, Jurgen Schüttler, Helmut Schwilden, Pablo Sepulveda, Frédérique Servin, Pamela Shafer, Steve Shafer, Stefan Shraag, Robert Sneyd, Robert Stoelting, Michel Struys, Nick Sutcliffe, and Jaap Vuyk.

From: Professor Steven Shafer.

Re: Open TCI Initiative, March 3, 2008.

Dear Colleagues:

My apologies for the tardy follow up to our one (and hopefully only) meeting of the Open TCI Initiative in CapeTown, South Africa. Most of you were able to attend the meeting in person. Those who were unable to attend were there ‘in spirit’, and I appreciate your support of the Open TCI initiative.

I want to again thank Fresenius Kabi for arranging the meeting and paying for the venue.

I don’t think it is worthwhile to have a point-by-point ‘minutes’ of the meeting. Instead, I’d like to focus on the meeting outcome.

Overview of the Open TCI Initiative

• It will function like ‘Open Source’ software.
• The data, software, support, and participation will be fully transparent.
• All data and software will be publically available.
• The Open TCI initiative will attempt to resolve the problems that plague existing TCI systems:
  • How should patients at the extremes of age and weight be handled
  • Existing ‘James Equations’ do not handle very obese individuals correctly.
  • Most pharmacokinetic data sets apply to adults, not children.
  • There are varying descriptions of the time course of the onset of drug effect.
  • As suggested by Struys et al.,[1] it may be the case that none of the available models accurately describes the time course in the first few minutes.
  • There are two different methods of implementing the effect site model: constant ke0 and constant tpeak. These produce slightly different results.
  • Users are expected to select from various pharmacokinetic and pharmacodynamic model. This is unacceptable. Users should be able to select a drug, and then enter covariates.
  • Device manufacturers should not have to decide what model is the ‘right’ model. This must come from a consensus within academic community.
  • Device manufacturers cannot change the pharmacokinetics of a pump readily. Changes in software, including changes in pharmacokinetic parameters, are prohibitively expensive. The academic community must provide guidance that device manufacturers can expect to perform well indefinitely.

Open TCI Principles

• Anyone can contribute data
• Anyone can download all data and analyses.
• If there are publications that come from the Open TCI initiative, those who have contributed to the effort will share authorship of the publications.
• All conclusions will be based on data.
• Any disagreements will be settled by data.
• Arguments from first principles will not be accepted unless there are no data on which to base conclusions.
• There will be 100% consensus among the participants supporting the final recommendations of the Open TCI Initiative.
• The results will be freely available to all companies interested in developing TCI, as well as all regulatory agencies looking to understand the basis of the recommendations from the Open TCI Initiative.
• All participants must agree to full disclosure of:
  • Consulting relationships
  • Confidentiality agreements
  • Honoraria
  • Institutional Support
  • Royalties, Stock
  • Other relationships
  • While financial details can be confidential, the nature of agreements (e.g., whether they are sources of revenue) should be disclosed.

The ‘Home’ of the Open TCI Initiative

• Currently, the Open TCI Initiative does not have a home within a recognized society.
• Various options were considered, including the International Anesthesia Research Society, the International Society of Anaesthetic Pharmacology, WorldSIVA, EuroSIVA, UKSIVA, the European Society of Anaesthesiology, and the Anesthesia Patient Safety Foundation.
• It was agreed that the International Anesthesia Research Society and the European Society of Anaesthesiology should be consulted to see if there was interest in being a home to the Open TCI Initiative.

Data sought by the Open TCI Initiative

• Propofol and remifentanil are the initial drugs of interest, although other drugs may eventually be analyzed by the initiative.
• Need as much dose vs. concentration data as possible, with particular emphasis on extremes of age and weight.
• Need as much high resolution measure of drug effect as possible, ideally accompanied by either dose information or drug concentrations.
• Need PK and PD data on propofol and remifentanil in various ethnic groups.
• Need PK and PD data on propofol and remifentanil in various disease states.
• Clinical effects (e.g., time of loss of consciousness) should be included when possible.
• It is likely that there is a great deal of data ‘in the middle’:
  • Patients of typical age and weight.
  • Patients at more or less steady state dosing
  • The critical need is not for the data ‘in the middle’, but the data at the extremes of age, weight, and arterial and venous concentrations, and measures of drug effect, within the first few minutes of dose administration.
  • Need to determine whether these data exist, or new studies are needed to inform the analysis at extremes of age, weight, and early time points.

TCI Analysis

• At present, the analysis will likely be done by Drs. Minto, Schnider, and Shafer.
• Others who wish to analyze the data, either in partnership with Drs. Minto, Schnider, and Shafer, or on their own, are welcome to do so.
• The ‘success’ of the Open TCI Initiative depends on vested individuals looking at the data independently and coming to congruent conclusions.
• We need regular progress reports to see that the initiative is actually moving forward.
• The Open TCI Initiative is an experiment.
• Need to assess if the experiment is working.
• Need a model that incorporates both arterial and venous samples.

TCI Results

• No data, no blah blah blah!
  • Do I have that right, Frédérique?
• We need a process to validate that data used in developing models.
• Not all data are equal.
• Need a rationale to explain why some data sets were included, and some were not included.
• Need to validate any models for the ages and weights of the data on which the model is based.
• In other words, should set limits (define ‘edge of risk’) on the weight and age for which we consider the model ‘validated’ by data.
• Should develop a plan for prospective validation:
  • Do we need a clinical trial to assess model performance? The answer, for now, is ‘maybe’.
• May be a role for cross validation.
  • Should the validation be against clinical measures (e.g., loss of consciousness), as well as blood concentrations and EEG measures of drug effect?

Next Steps

• Project Management
• Steve Shafer will direct the project (until someone else volunteers to do so)
• Project team will consist of:
  • Tom De Smet in collaboration with other engineers
  • Charles Minto will direct the data modeling
  • Each Society will ask for a representative, who will actually do some of the work
• Prepare the Open TCI Website
  • Need registration page
  • Anyone can register
  • Need data upload page
  • Anyone can read.
  • Only registered participants can upload data.
  • Data, once uploaded, cannot be deleted (for auditability).
  • Revised data sets can be uploaded, with a different name for the data file.
  • Excel will be the common format for uploaded data.
• Need participant disclosure page.
  • Everyone can read the disclosure pages.
  • Only registered users can update their information.
  • Obtain data inventory.
  • Individual participants should indicate what data they are willing to upload.
  • For now, we just want an inventory of data, not the data themselves.

I believe that covers our discussions. I would appreciate all participants in the first (and hopefully last) meeting of the Open TCI Initiative to review those data that you are willing to contribute to the initiative.

Thank you again for your support of the Open TCI Initiative.

Sincerely,
Steven Shafer

Editor-in-Chief, Anesthesia & Analgesia
Professor of Anesthesiology, Columbia University
Adjunct Professor of Anesthesia, Stanford University
Adjunct Professor of Biopharmaceutical Sciences, UCSF