Fri, 20 May 2005
ARMA was a government-sponsored, $41 million experiment involving 861 critically ill patients with severe lung injury–i.e., acute respiratory distress syndrome (ARDS). It has been at the center of continuing debate prompted by revelations that researchers at 14 of the nation’s major medical research institutions comprising the ARDS Network,* violated fundamental medical ethics principles.  
Patients with ARDS almost always need to be supported with mechanical ventilation to help them breathe. During this kind of support, breaths, termed tidal volumes, are administered to patients with a ventilator to inflate the lung to deliver oxygen and remove carbon dioxide from the blood.
However, lungs with ARDS are much stiffer than healthy lungs. In the 1980’s emerging research strongly suggested that although necessary for survival, if tidal volumes with mechanical ventilation were too large, the high airway pressures they produced might themselves be harmful.
In 1992, a survey of more than 1,000 physicians caring for patients with ARDS showed that most were aware that tidal volumes producing increased airway pressures might be harmful and that physicians adjusted their practice to account for this.  A consensus conference of physicians in 1993 warned of the need to reduce tidal volumes if airway pressures were high, and recommended that physicians reduce tidal volumes when necessary to ensure that airway pressures remained below a potentially harmful level. 
NIH sponsored the ARMA study to investigate the benefits of reducing tidal volumes in ARDS patients. Between 1996 and 1999, the ARDS Network enrolled 861 patients out of an eligible pool of 3,448 patients. Patients who were enrolled in this experiment had been receiving oxygen at a tidal volume setting according to the judgment of their primary, critical care physician. They were withdrawn from that ventilation setting, and randomized to receive one of two tidal volume settings–either high (12 ml/kg) which was termed “traditional” or low (6 ml/kg).
The experiment failed to include a comparison control group receiving the current best treatment–which included individualized ventilation settings determined by the condition of the individual patient’s lungs.
Most of the patients enrolled in this randomized experiment were unable to provide informed consent. Furthermore, the consent forms that had been approved by the institutional review boards (IRBs) at the participating hospitals ALL failed to comply with federal requirements of disclosure.
In fact, none of the consent forms informed patients or their families that they would be withdrawn from current best therapeutic treatment and randomly assigned to a fixed ventilation setting which might increase their risk of dying. Neither were they informed that an alternative to participating in the experiment was available at the same medical center.
On July 29, 2002, the Alliance for Human Research Subjects filed a complaint with the federal Office of Human Research Protections about violations of informed consent. See: OHRP response to the AHRP complaint.
After enrolling more than 860 patients the study was stopped because mortality was substantially higher in the high tidal volume group (12ml/kg, 41%) compared to the low tidal volume group (6ml/kg, 31%). The ARDS investigators concluded that the low tidal volume (6ml/kg) increased survival rate and recommended that it should be applied to all patients with ARDS. 
Following publication of the trial in the New England Journal of Medicine, critical care physician-scientists at the NIH who analyzed the ARDS Network report raised concerns about the design of the experiment in the American Journal of Respiratory and Critical Care Medicine (AJRCCM). 
This independent analysis raised the following concerns about the design of the trial: First, airway pressures associated with the 12 mg/kg tidal volume were significantly higher than what critical care physicians were routinely administering in standard practice to their patients at the time of the trial. Second, the trials failed to include a control group in which patients would be provided individualized ventilation settings–as is the practice in standard care, and is required under the Declaration of Helsinki ethical research standards.  Ethical research requires that:
“The benefits, risks, burdens and effectiveness of a new method should be tested against those of the best current prophylactic, diagnostic, and therapeutic methods.”
Other groups of investigators were also examining what routine ventilatory care was both before and during the ARDS trial. These studies showed that physicians varied tidal volumes based on how injured or stiff the lungs of patients were. A very sick patient with a very stiff lung would get a low tidal volume while a patient with a more pliant lung would receive a higher tidal volume.
Randomizing patients to high or low tidal volumes without regard for their underlying level of lung injury was (and is) contrary to what was at the time–and continues to be–the best therapeutic method of care.
Therefore, the significant findings of the ARDS Network experiment may relate largely to having increased mortality in patients with very stiff lungs who had their tidal volumes increased to 12ml/kg.
Ultimately the ARDSNET trial is impossible to interpret and its recommendations that all patients with ARDS should get very low tidal volumes–should be rejected precisely because the trial failed to provide a randomly-selected control group of patients who received standard treatment.
“Mechanical ventilation in ARDS: One size does not fit all,” an editorial by Katherine Deans, Peter Minneci, Xizhounog Cui, Steven Banks, Charles Natanson and Peter Eichacker, in the current issue of the AJRCCM re-analyzes the ARDSNET data and additional data obtained via the Freedom of Information Act. 
The editorial analysis shows that “control” patients receiving high tidal volumes had a mortality rate (41%) higher than those eligible patients not enrolled in the trial (32%).
Deans, et al, examined data from patients before and after randomization. With enrollment in the trial, earlier routine care for patients was stopped. Analysis showed that in patients with stiff (sicker) lungs, increasing tidal volumes from routine care levels after randomization increased mortality rates. Similarly, in patients with less injured, more elastic lungs, reducing tidal volumes from routine care also increased mortality
If lung stiffness and the resultant airway pressures dictated tidal volume selection by physicians routinely, then patients who were randomized to receive tidal volumes different from those their primary care physicians had prescribed were put in harm’s way.
The lessons from the ARMA-ARDS Network experiment:
If the researchers had complied with ethical standards of informed consent and the Helsinki provision requiring that an experimental treatment must be tested against “the best current prophylactic, diagnostic, and therapeutic methods” lives would not have been sacrificed needlessly. Furthermore, information might have been gained to benefit future patients as well.
But given the ARDS Network’s disregard for the patients’ human right to informed consent, and their flawed trial design, the $41 million ARMA experiment was an egregious mistrial which increased mortality.
A question arises about the ethical justification of randomization to a fixed dose treatment in research involving critically ill patients whose very life depends on sensitive adjustments to every aspect of their care and treatment from moment to moment.
Why should investigators be free to use procedures that increase the risk of death when the information they seek may be obtained using other research avenues?
Indeed, the second principle of the Nuremberg Code (1947) stipulates specifically that: “The experiment should be such as to yield fruitful results for the good of society, unprocurable by other methods or means of study, and not random and unnecessary in nature.” 
Finally, the ARMA-ARDS experiment is an indictment of the IRB system and the bioethics community. IRBs at major medical research institutions that approved this experiment and others like it have demonstrated their incapacity to exercise their legal authority for the protection of human subjects. By approving experiments whose very design violated fundamental ethical principles and whose patient enrollment failed to comply with federal informed consent requirements, they share in the responsibility for the resulting preventable deaths.
- “Ventilation with Lower Tital Volumes as Compared with Traditional Tidal Volumes for Acute Lung Injury and the Acute Respiratory Distress Syndrome,” New England Journal of Medicine (2000, May 4), 342 (18): 1301-1308.
Carmichael LC, Dorinsky PM, Higgins SB, Bernard GR, Dupont WD, Swindell B, et al. Diagnosis and treatment of acute respiratory distress syndrome in adults: an international survey. J Crit Care 1996; 11(1):9-18.
Slutsky AS. Mechanical Ventilation. American College of Chest Physicians’ Consensus Conference. Chest 1993; 104(6):1833-59.
Eichacker, et al, “A Metaanalysis of ALI and ARDS Trials Testing Low Tidal Volumes.”
WORLD MEDICAL ASSOCIATION DECLARATION OF HELSINKI
“Mechanical ventilation in ARDS: One size does not fit all.” Editorial by Katherine Deans, Peter Minneci, Xizhounog Cui, Steven Banks, Charles Natanson and Peter Eichacker. Crit Care Med 2005 Vol. 33, No. 5, pp. 1141-1143.
THE NUREMBERG CODE [from Trials of War Criminals before the Nuremberg Military Tribunals under Control Council Law No. 10. Nuremberg, October 1946-April 1949. Washington, D.C.: U.S. G.P.O, 1949-1953.] http://www.ushmm.org/research/doctors/Nuremberg_Code.htm
**ARDS Network participants:
The Cleveland Clinic Foundation and MetroHealth Medical Center, Cleveland; Duke University Hospital and Veterans Affairs Hospital, Durham, NC; LDS Hospital and University of Utah Hospital, Salt Lake City; McKay-Dee Hospital, Ogden, UT; University of California Medical Center, Moffitt-Long Hospital, and San Francisco General Hospital, San Francisco; University of Colorado Health Sciences Center, Denver Health and Hospitals, Veterans Affairs Hospital, and Rose Medical Center, Denver; University of Maryland Hospital, Veterans Affairs Medical Center, Johns Hopkins-Bayview Medical Center, and Johns Hopkins Hospital, Baltimore; University of Michigan Medical Center, Ann Arbor; Hospital of the University of Pennsylvania and Thomas Jefferson University Hospital, Philadelphia; University of Washington / Harborview Medical Center, Seattle; and Vanderbilt University Hospital, Nashville. The Coordinating Center is at Massachusetts General Hospital, Boston. See: http://www.ardsnet.org/links.php
Contact: Vera Hassner Sharav
Editorial #050090 Crit Care Med 2005 Vol. 33, No. 5
Mechanical ventilation in ARDS: one size does not fit all*
Katherine J. Deans, MD 1,2, Peter C. Minneci, MD 1,2, Xizhong Cui, MD1, Steven M. Banks, PhD 1, Charles Natanson, MD 1, and Peter Q. Eichacker, MD 1.
In this issue of Critical Care Medicine, Kallet et al. report a significant improvement in mortality in patients with adult respiratory distress syndrome (ARDS) and acute lung injury (ALI) who received lung protective ventilation based on the recommendations of theARDSNetwork trial of low tidal volume ventilation (ARMA) [6 mL/kg predicted body weight (PBW)] compared to historical controls receiving routine care (32% vs. 51% mortality, respectively; p<0.004). (1, 2) Similar to the ARMA trial, Kallet et al. conclude that despite the heterogeneity of underlying lung injury typically present in this patient population, their findings “strongly suggest that the low VT strategy may benefit all patients with ALI or ARDS”. In contrast, other leaders in the field believe that tidal volumes in patients with ALI and ARDS should be titrated based on surrogate markers of lung injury such as airway pressures or lung strain.(3-7) These physicians indicate that although some patients, such as those with poor pulmonary compliance and high airway pressures, would benefit from very low tidal volume, others with less severe lung injury may require larger volumes to maintain ventilation and avoid alveolar collapse.
Kallet et al conclude that ventilation with low tidal volumes in all patients with ALI and ARDS is superior to the routine care employed at the time of their historical control. (1) Before adopting these recommendations, it is important to assess the consistency of these results with other available data comparing low tidal volume ventilation with contemporaneous routine practice. Because the historical control group reported by Kallet et al received routine care at a participating ARDSNetwork center during a period of time (1/98 to 12/99) overlapping the ARMA trial (3/96 to 3/99) (1, 2), one would expect that survival rates in low tidal volume patients in the ARMA trial would be both (1) improved compared other patients receiving routine care at ARDSNetwork centers during the time of the trial (2) and uniformly better than control patients in the ARMA trial independent of their underlying level of lung injury. To investigate these questions, ARMA trial data obtained from the Office of Human Research Protections via the Freedom of Information Act was analyzed with approval from the Internal Review Board at the Clinical Center at the National Institutes of Health. (8)
The ARDSNetwork screened over 6,000 patients for enrollment in the ARMA trial. (2, 8) There were 2,587 patients who met enrollment criteria, but were ineligible for technical reasons (Figure 1). This ineligible group of patients received routine care during the course of the trial and their mortality rate was documented at intervals similar to study participants. Therefore, this group of ineligible patients provides an interesting comparator group reflective of current practice at the time of the trial, similar to the historical control group reported by Kallet et al. Interestingly, these ineligible patients receiving routine care maintained a mortality rate comparable to the low tidal volume arm of the ARMA trial (31.7% vs. 31% respectively). (2, 8) This is not consistent with the findings of Kallet et al. in which the historical control group mortality was significantly higher than the low tidal volume arm (51% vs. 32% mortality respectively; p<0.004).
Given that the mortality rates of the low tidal volume groups were consistent (31% vs. 32%), the discrepancy in results between the ARMA trial data and that of Kallet et al may be explained by the difference in control mortality rate (31.7% vs. 51% respectively). Both the historical control group from Kallet et al and the observational cohort from the ARMA trial are susceptible to numerous biases. However, the cohort from the ARMA trial was treated contemporaneously by the same physicians at the same hospitals during the time of ARMA trial, whereas Kallet et al employed a historical control.
Despite attempts by the authors to control for differences between the two groups, the historical controls used in the study by Kallet et al. had significantly higher APACHE scores than the low tidal volume strategy treatment group indicating that they may have been a more severely ill group of patients. In the absence of a randomized controlled trial which directly studies routine care compared to a low tidal volume ventilation strategy, the observational cohort from the ARMA trial may provide a more robust comparator group on which to make observations about routine care versus low tidal volume ventilation.
Consistent with the reported results of the ARMA trial, Kallet et al suggest that the 6 mL/kg PBW ventilation confers a survival advantage over control which is solely based upon the lowering of tidal volume.(1, 2) Although Kallet et al. do not provide comparative data on compliance, our analysis of the ARMA trial data reveals that outcome in the trial depended on baseline pulmonary compliance. Examination of the pre-randomization compliance of patients in the ARMA trial revealed that prior to randomization, tidal volume was titrated based on pulmonary compliance (Figure 2). (8) Interestingly, the mortality effect of changing tidal volume was significantly associated with pre-randomization lung compliance (p=0.003 for interaction) (Figure 3). (8) Using logistic regression to correct for age, APACHE II score, and FiO2/PaO2 ratio differences, individually or combined, the p-value for this interaction was essentially unchanged. Patients with more compliant lungs did poorly if tidal volumes were lowered. In contrast, those with relatively less compliant lungs did well if tidal volumes were lowered. Identification of this interaction indicates that the overall survival benefit reported in the ARMA trial is not applicable to all ARDS/ALI patients. This finding is consistent with the philosophy that parameters of mechanical ventilation should be titrated based on surrogates of disease severity. (3-7)
In light of the unexplained high control mortality rate reported in the trial by Kallet et al, and the marked relationship between compliance and mortality in the low tidal volume arm of the ARMA trial, a single ventilatory strategy such as 6 mL/kg tidal volume ventilation may not be ideal for all patients with ARDS and ALI.
1. Kallet RH, Jasmer RM, Pittet JF, et al.: Clinical implementation of the ARDS-Network protocol is assoicated with reduced hospital mortality compared to historical controls. Crit Care Med 2005;
2. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network. N Engl J Med 2000;18:1301-8
3. Gattinoni L, Caironi P and Carlesso E: How to ventilate patients with acute lung injury and acute respiratory distress syndrome. Curr Opin Crit Care 2005;1:69-76
4. Marini JJ and Gattinoni L: Ventilatory management of acute respiratory distress syndrome: a consensus of two. Crit Care Med 2004;1:250-5
5. Petrucci N and Iacovelli W: Ventilation with smaller tidal volumes: a quantitative systematic review of randomized controlled trials. Anesth Analg 2004;1:193-200
6. Tobin MJ: Culmination of an era in research on the acute respiratory distress syndrome. N Engl J Med 2000;18:1360-1
7. Ricard JD: Are we really reducing tidal volume–and should we? Am J Respir Crit Care Med 2003;10:1297-8
8. Data provided to the Office of Human Research Protections from ARDSNetwork investigators from the ARMA trial for use at the June 9 – 11, 2003 consultants meeting. Available under the Freedom of Information Act. Requests should be sent to Public Health Service, Freedom of Information Office, Rm 17A46, 5600 Fishers Lane, Rockville, MD 20857.