Acute Respiratory Distress Syndrome – ARDSNet experiment criticized – OHRP

Acute Respiratory Distress Syndrome – ARDSNet experiment criticized – OHRP

Fri, 28 Feb 2003

An Acute Respiratory Distress Syndrome (ARDS) study conducted between 1996 and 1999, on critically ill patients at 20 medical facilities (14 major research centers belonging to the ARDS Network), was published in The New England Journal of Medicine (NEJM) in 2000. The study compared two extreme air-pressure ventilation treatment strategies; it was halted when the investigators claimed that data gathered from 861 patient/subjects demonstrated a 22% reduction in the death rate of those treated with an experimental low-pressure air ventilation compared to others treated with high-pressure. This government sponsored study is under a reopened federal investigation, and its claimed findings are disputed.

In its initial investigation, the Office of Human Research Protections (OHRP) faulted all the participating research centers for violations of informed consent–which was mostly obtained from surrogates. OHRP also found that risks had not been adequately disclosed, in some cases OHRP faulted the institutionally approved Informed Consent documents for failing to inform patients and / or surrogates about the risks and discomfort of low volume air ventilation that the investigators endorsed–i.e., patients became “very dyspneic and agitated.” [See: http://ohrp.osophs.dhhs.gov/detrm_letrs/YR02/feb02d.pdf See also, http://www.ahrp.org/ethical/CaromeARDS082802.php]

The ARDS study has raised serious ethical concerns about the research review and approval process in multi-center studies at prominent institutions that receive hundreds of millions of dollars in public funds. Complaints about the study design, the absence of a control group, violations of informed consent, failure to conduct a pilot study prior to this large-scale study, and myriad other troubling aspects have resulted in a heated controversy and the suspension by NIH of a companion ARDSNetwork study.

Following a complaint filed by the Alliance for Human Research Protection (AHRP) (July 29, 2002) and a critical analysis by senior researchers* at the NIH, OHRP re-opened its investigation. *[See: Metaanalysis of ALI and ARDS Trial Testing Low Tidal Volume, by Drs. Peter Eichacker, Eric Gesrtenberge, Steven Banks, Xizhong Cui, and Charles Natanson, in the American Journal of Respiratory and Critical Care Medicine On August 28, 2002,166:1510-1514].

The ARDSNet investigators defended the study in their letter to the journal [See: Brower RG, Mathay M, and Shoenfeld D. Am J Respir Crit Care Med 2002, 166:1515-1516]

Two letters in the current issue of the same journal raise further serious concerns: Dr Christopher Parshuram and Dr. Brian Kavanagh of the University of Toronto point out a fallacy in the ARDS study approach: “significant differences between groups in a randomized trial inform only about the relevant merits of the tested therapies, and in isolation cannot differentiate between more benefit in the ‘superior’ group versus more harm in the ‘inferior’ group.”

Drs. Eichacker, Banks and Natanson analyzed additional data (provided by Bower et al, in their letter). The new data was obtained from critically ill patients who declined to participate in the ARDSNetwork trials. After laying out all the patient data on a graph, Eichekar et al concluded: “It is clear that the tidal volumes received by patients before randomization were normally distributed in a bell shaped curve.” But after being assigned to the “control” group and their ventilation increased according to the protocol, the death rate climbed.

“Control patients (n=429) had a significantly higher mortality rate compared with these eligible nonparticipants (40 versus 31%, p = 0.02). Importantly, the low mortality rate in the nonparticipants (n=260)is equivalent to the mortality that was reported in ARDS patients in 1996, the year the ARDSNetwork trial began.”

The ARDSNetwork investigators themselves concluded in their letter: “…critically ill patients who participated in this trial did not fare better overall than eligible patients who declined to participate.”

Note our tabulation of the results:

Trial Arm N Deaths (%)

High volume 429 40 Low volume 432 31 Refusers 260 31

There is a 9 percent difference in death rates between patients in the high volume arm and both in the low volume arm and in patients who directly refused or whose proxies refused participation.

Eichaker et al, point out: “This trial compared approximately the 3rd to the 80th percentile of current practice without assessing the most commonly used level of care.” For that reason, they say, “the ARDSNetwork trial does not appear to answer the most important question: whether the experimental (low volume of air at 6ml / kg) was better, worse or no different than routine care given to most patients with ARDS.”

On October 7, 2002, OHRP wrote a highly critical 29 page letter to the ARDSNetwork, stating “OHRP continues to have serious unresolved concerns…trials failed to comply with key requirements of HHS regulations” OHRP requested the institutions to investigate further, requesting full patient chart information from all trial centers by December 6, 2002.

The OHRP letter can be accessed on the AHRP website at: http://www.ahrp.org/ARDSpages/ARDSindex.php

The claim has been made that patients in clinical trials receive better care than those in standard care. Is that claim credible?

The ARDSNet study demonstrates the pitfalls of testing a “theory” on severely ill patients, in a large-scale experiment, without first analyzing “the range and frequency of current practice under question.” Patients were put at increased risk in the ARDS study which “incorporated two study arms, neither of which represented predominant physician practice.” This experiment demonstrates that patients are at far greater risk of death in poorly designed clinical trials than if they receive standard care.

But where were the oversight committees (i.e., IRBs) that are authorized to protect the safety of research subjects? Medical research ethics requires that an experimental an treatment or method must be tested against “the best current prophylactic, diagnostic, and therapeutic methods.” [See Declaration of Helsinki]

Did no one at any of the 20 (or so) IRBs raise concerns about approving an experiment without a control group receiving “standard of care?” Did no one at these major research institutions suggest first doing a pilot study to make sure the proposed “theory” is valid?

What is more, how could IRBs and the National Heart Lung and Blood Institute have approved a consent form and informed consent procedure that failed to alert the patients or their proxies of the possibility that participation in the research might INCREASE the risk of death?

Is it a reasonable assumption that the provision of such vital information would have made it more difficult to recruit 861 subjects—-the exact number that precisely satisfied the sample size to achieve the .85 statistical power of the test? Was this statistical power needed to impress peer reviewers of the merit of the study’s findings at journals such as the New England Journal of Medicine?

OHRP’s letter states: “there appears to be insufficient evidence to support any conclusion that mechanical ventilation management with low tidal volume intervention (6ml/kg)is superior to individualized ‘standard of care’.” [See: http://www.ahrp.org/ARDSpages/ARDS02.html

It is deeply troubling that the NEJM has not yet issued a retraction of the questionable claim of findings which may mislead physicians in intensive care units.

How many other published clinical trials have been conducted with greater attention paid to satisfy a statistically significant finding than to patient safety?

For documents detailing the ARDS controversy visit the AHRP website at: www.ahrp.org