Critical Care Plus

ARDSNetwork Mechanical Ventilation Trial Still Making Waves

Clinicians adopting study recommendations, but controversy continues

By Julie Crawshaw, Critical Care Plus Editor

The ARDSNetwork trial that began in 1996 (ARMA trial)1 was designed to test the validity of using lower tidal volumes in mechanical ventilators was halted because critics who were not a part of the ARDSNetwork argued that the control arm chosen was harmful and unethical.

The study ended in 1999 but the controversy continues today. Early study data indicated that the lower tidal volume was effective. And since that data was released, low tidal volume ventilation has been embraced by clinicians around the world, dramatically changed ICU practice, and improved survival rates in ARDS patients. Yet critics continue to argue that study results can’t be correctly interpreted because all that can be concluded is that the low tidal volume arm was better than a harmful control arm.

If you believe the 22% reduction in the death rate of ARDS patient on low tidal volume occurred only because the protocol was measured against something harmful, then it’s not very impressive, notes Gordon D. Rubenfeld, MD, of the Division of Pulmonology & Critical Care Medicine at Harborview Medical Center in Seattle.

But Rubenfeld believes it is wrong to say that the control arm strategy used in the study deviated considerably from the standard of care. He says that many doctors use the 12 mL/kg Predicted Body Weight (PBW) tidal volume that served as the study’s control arm rather than the 6 mL/kg study results showed improves survival.

Misunderstanding Leads to Criticism

Much of the criticism arose and continues, Rubenfeld says, because many doctors mistakenly believed the 12 mL/kg parameter referenced Measured Body Weight (MBW) which is approximately 20% less than PBW. "It’s true that the only thing we know is that 6 mL/kg of predicted body weight is better than 12 mL/kg. We don’t know that 6 is better than 4, 7, 8, 9 or 10 mL/kg." Rubenfeld says. "It’s also true to say we’re never going to know those things because it would take thousands of patients—nobody’s going to pay for that."

Rubenfeld says that much work remains before the 6 mL/kg tidal volume is universally accepted. It’s a best practices issue, Rubenfeld believes. Nurses and respiratory therapists have problems using 6 mL/kg because patients may appear less comfortable than they do with 12 mL/kg, Caregivers have to get used to a whole new set of laboratory tests because patients tend to drop their oxygen saturation.

Furthermore, some data show that primary physicians recognize ARDS in only about 30% of their patients who have it. "If you don’t recognize that a patient has acute lung injury you’re not going to use the protocol," Rubenfeld notes.

Two physicians who are not members of the ARDSNetwork complained to the federal Office of Human Research Protections (OHRP), which in turn conveyed its concerns to the National Heart, Lung and Blood Institute (NHLBI), which in its turn sponsored an ongoing trial in which patients were receiving 6 mL/kg PBW. This trial had certain design characteristics that were similar to the tidal volume trial. The NHLBI voluntarily suspended the trial while the ARDSNetwork investigators studied the issues and prepared responses prior to submitting their responses last June.

Everybody has read the criticisms of the study, Rubenfeld adds, but few in the front lines realize that two independent review boards (IRBs), the first convened by the National Institute of Health and the second by OHRP, reviewed the ARDSNetwork trial and found it to be valid.

"Many practitioners are left with a vague sense that something was wrong with the study, so they continue doing what they’ve been doing," Rubenfeld says.

Problems with uptake of evidence are hardly unique to critical care, as Rubenfeld points out. But practitioners in other specialties are ahead of critical caregivers in encouraging and reminding practitioners that they need to adhere to the best practices evidence demonstrates. "What’s new in critical care is that now we actually have evidence," Rubenfeld says. "Ten years ago we didn’t have many clinical trials we could use as evidence—there were lots of evidence in cardiology and oncology—but not in critical care."

In most of those fields thoroughly studied, the gap between evidence-based medicine and practice is lower because specialty societies have pushed for and gotten programs that encourage practitioners to use evidence-based guidelines. Rubenfeld stresses the need for such programs in critical care, adding that this is definitely a challenge for big critical care professional organizations.

Unfortunately, critical care has comparatively few good clinical trials from which to distill best-practices information. "We need to invest in more projects such as the randomized controlled trial in 10 community hospitals we’re running here to try and change ventilator behavior at the bedside," Rubenfeld says.

Lack of IRB Information Cited

Patricia El-Hinnawy of the OHRP says her organization found that when reviewing and approving the ARMA trial, the reviewing IRBs failed to receive or request sufficient information to make the required determinations.

El-Hinnawy reports no estimated date for completion of review. "There are some exchanges of information and analyses from the ARDS Net researchers that our office needs to look at before we will consider the case to be closed," she says.

A 15-page letter sent by OHRP to concerned parties last July said that "almost all of the consultants engaged by OHRP opined that risks to subjects participating in the ARMA trial were minimized and reasonable in relation to anticipated benefits to the subjects and the importance of the knowledge that was expected to result."2

The letter continued, "OHRP believes, however, that the interests of future human subjects would be served best by further discussion within the scientific and bioethics communities about issues regarding appropriate research design in the absence of a standard of care that have been raised in the context of OHRP’s compliance oversight evaluation of the ARMA trial. OHRP encourages such discussions."

Survey of Tidal Volumes Used Preceded Study

According to B. Taylor Thompson, MD, director of the Medical Intensive Care Unit and Medical Director of the ARDSNetwork Clinical Coordinating Center at Massachusetts General Hospital in Boston, ARDSNetwork investigators relied on a survey of ICU practitioners and an analysis of tidal volumes used in clinical care when they chose the study parameters.

Taylor adds that this analysis showed that 10 mL/kg PBW was commonly preferred and used by clinicians, though tidal volumes of 10-15 mL/kg MW had been recommended by experts in the field for many years.

"Accordingly, this tidal volume was a reasonable representation of both usual practice in the early 1990s and the traditionally recommended approach," Taylor says. "We chose PBW, based on height and gender, because height and gender better estimate an individual patient’s lung size."

Those clinicians who participated in the trial had used a wide range of tidal volumes prior to enrolling in the trial. "The mean tidal volume was 10.4 mL/kg PBW, which is 1.4 mL/kg lower that our traditional group," Taylor says, "(That’s) a very small difference in tidal volume of unclear clinical importance even in retrospect."

He points out that by the time of the tidal volume trial clinicians had already begun moving from traditionally recommended tidal volumes to lower ones, as one consensus recommendation suggested clinicians do. However, this recommendation was not based on sound clinical evidence and it was unclear, if indeed clinicians’ practices were indeed changing, that they were changing in the right direction.

Another explanation for the slightly lower tidal volumes in use at ARDSNetwork sites, as Taylor points out, is that during the trial routine care was influenced by the conduct of the trial. Clinicians at study hospitals had been informed of the study and its rationale. Knowing that researchers were evaluating tidal volumes as low as 6 mL/kg may have influenced clinicians to modify their practice and to choose tidal volumes intermediate between the study groups—a well-known phenomenon called the Hawthorne effect. Thus, as Taylor observes, baseline data in any clinical trial may not be an accurate source of information about usual care prior to a trial.

Earlier Trials Used Lower Volumes

Taylor adds that in two large clinical trials that included ARDS patients during the early 1990s, physicians prescribed tidal volumes of approximately 10.5-11.5 MBW. In the ARDSNetwork trial, MBW exceeded PBW by 20% on average. Assuming that this 20% difference was true in the early 1990s-and Taylor observes there’s no reason to think otherwise-then physicians were prescribing tidal volumes of 12.5-13.5 mL/kg PBW in the early 1990s. Taylor adds that in another recent trial MBW exceeded PBW by more than 30%. "If these things were so, then it would seem that the NIH’s quarrel with the study is not with improved outcomes associated with using a 6 mL/kg perceived body weight, especially as early data indicated low tidal volume was effective," Taylor says.

"On balance, the ARDS Network trials have withstood unprecedented external review and were found to be soundly designed and important," Taylor says. "We look forward to continuing our trials with the hope that we can provide ICU clinicians with sound clinical evidence to improve the care of critically ill patients around the world."

For more information contact Gordon Rubenfeld, MD, at (206) 731-8584, B. Taylor Thompson, MD, at (617) 724-3705, and Patricia El-Hinnawy (301) 435-5654.


1. Prospective, randomized, multi-center trial of 12 mL/kg vs 6 mL/kg tidal volume positive pressure ventilation for treatment of acute lung injury and acute respiratory distress syndrome.

2. Letter sent by the OHRP July 3, 2003, to all ARDS Network Institutions.