Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Effect of enteral feeding with eicosapentaenoic acid, gamma-linolenic acid, and antioxidants in patients with acute respiratory distress syndrome

Enteral Nutrition in ARDS Study Group. Gadek JE, DeMichele SJ, Karlstad MD, et al.

Crit Care Med 1999;27:1409-20. [abstract]

Reviewed by Scot Bateman MD, Children's Hospital, Boston

Review posted December 4, 1999

 

I. What is being studied?:

The study objective:

The objective was to determine if enteral nutrition with EPA (eicosapentaenoic acid) + GLA (gamma-linolenic acid) - to reduce pro-inflammatory eicosanoid synthesis - and antioxidants may reduce pulmonary inflammation, improve oxygenation and clinical outcomes in patients with ARDS.

The study design:

Prospective, multi-centered, double-blind, controlled trial.

The patients included:

Adults with 1) illness associated with ARDS (acute bacterial, viral, or fungal pneumonia; sepsis syndrome; aspiration; inhalation of gaseous agent; trauma; burns). 2) bronchoalveolar lavage evidence of pulmonary inflammation [PMN>10%]. 3) PaO2/FiO2 < 250 torr but > 100 torr. 4) had enteral access providing a mechanism for either gastric, duodenal, or jejunal tube feeding. 5) Diffuse pulmonary infiltrates on CXR.

The patients excluded:

a) left ventricular failure, b) lung cancer, c) hematological malignancy, d) acute gastrointestinal bleeding, e) head trauma, f) stroke or subarachnoid hemorrhage, g) severe immunodeficiency, h) use of steroids or nonsteroidal anti-inflammatory within last 24 hours, i) HIV, j) pregnant.

The interventions compared:

Patients were randomized to receive either a) control diet (ready-to-feed high-fat, low-carbohydrate enteral nutrition formula providing complete, balanced nutrition) vs. b) isocaloric and isonitrogenous diet with addition of EPA+GLA lipids and antioxidant vitamins.

The outcomes evaluated:

Many outcomes were following in this study.

Once patients reached 75% of basal energy expenditure via enteral feeding, outcomes were assessed as

1. clinical and laboratory data- daily lab tests, plasma lipids, ventilator settings, gas exchange
2. bronchoalveolar lavage data- BAL cell counts
3. outcome variables- ventilator days, ICU days, time on supplemental oxygen, new organ failure, and mortality,

II. Are the results of the study valid?

1. Is there a strong, independent, consistent association between the surrogate end point and the clinical end point?

The clinically important outcome in ARDS has always been mortality. This has been the focus of most of the clinical trials in ARDS research. With historic mortality of ARDS around 50-60%, this was the obvious endpoint to target in studies. More recently the mortality rate in ARDS has improved down to approx. 30-40%. This has led to discussions of whether mortality should be the focus of clinical trials in ARDS (1, 2). Other post-ARDS morbidities, which may become important variables for end points need to be clearly defined. Recent study of ARDS survivor analysis (3) suggests more work needs to be done to better identify morbidities which should be targets of clinically relevant endpoints. Surrogate endpoints have thus become particularly prevalent in studies of treatment strategies in ARDS research. The paper chosen for review here uses an array of surrogate endpoints which merit evaluation; indirect measures of inflammation, oxygenation and ventilator measures, organ failure rates, and ventilator and ICU-free days. All of these endpoints make physiologic sense to practitioners as important parameters, which should influence the mortality rate.

Answering the critical questions for end point validity for each of these endpoints is hampered by very little randomized controlled studies showing improved outcome in ARDS. Therefore, we are left only answering the first question about the validity of the surrogate endpoints. Looking at correlations of various factors and mortality have been evaluated by looking at predictions of mortality. Two recent multivariate analyses studies (4, 5) of acute lung injury confirmed many earlier studies by failing to show any relationship between oxygenation, inflammation or ventilatory measures with mortality. The study by Doyle (5) did show an odds ratio of 8.1 (2.1-30.8) of mortality with the development of non-pulmonary organ system dysfunction, but it was in organ dysfunction which occurred prior to ICU admission and not that which developed in the ICU. A multiple organ dysfunction score has been used in critical care mortality prediction models and has validated prospectively (6). Organ failure and ARDS relationships are difficult to correlate, however, because many patients with ARDS do not get other organ failure. There does not seem to be any relationship with ICU or ventilator-free days with mortality in ARDS. Obviously, there is a cost improvement to the hospital for fewer ventilator days or ICU days. This outcome has become very prevalent in critically ill patient studies but clearly needs more study.

2. Is there evidence from randomized trials in other drug classes that improvement in the surrogate end point has consistently led to improvement in the target outcome?

No (see above).

3. Is there evidence from randomized trials in the same drug class that improvement in the surrogate end point has consistently led to improvement in the target outcome?

No (see above).

IIa. Validity Questions for Therapy Articles:

Primary questions:

1. Was the assignment of patients to treatments randomized?

Yes, from 5 centers, with a permuted block randomization design.

2. Were all patients who entered the trial properly accounted for and attributed at its conclusion?

Was followup complete? and were patients analyzed in the groups to which they were randomized?

No, but they did an intention to treat analysis that confirmed their main outcome results. 146 met criteria, but only 98 were evaluated and then only 65 were evaluated at day 7. Intent to treat analysis was done on 142 patients. Follow-up was complete for the 65 patients and there were no crossovers in the study.

Secondary questions:

3. Were patients, health workers, and study personnel "blind" to treatment?

Yes.

4. Were the groups similar at the start of the trial?

Yes. Age, type of illness, laboratory data and organ system failure at baseline were all similar between the two groups.

5. Aside from the experimental intervention, were the groups treated equally?

Unclear. Other aspects of the patents' care were not specified.

III. What were the results?

1. How large, precise, and lasting was the treatment effect? (Effect should be large, precise, and lasting to consider a surrogate trial as possible basis for offering patients the intervention.)

a. Total cell and neutrophil counts in BAL fluid: Interpreting the numbers from their graph they report a significant drop in the study group at 4 days. Total cells decreased from 782 to 432 +/- 150 in study vs. increased from 537 to 687 +/- 275 in control at 4 days. PMNÕs decreased 570 to 220 +/- 100 in the study group vs. increased 307 to 407 +/- 250 in the control group. P=0.008. The difference was not sustained beyond 4 days.

b. Gas exchange: P/F ratio in study group went from 156to 217 to 198 at baseline to day 4 to day 7 vs. the control groups whose P/F ratio went from 170 to 169 to 154. P=0.009. Oxygenation improvements were significant at 4 and 7 days suggesting a lasting effect. Other parameters showed: a) higher PIP and FiO2 in control at 4 days, not sustained at 7 days, b) lower PaO2 in control group at 7 days, and c) significantly lower minute ventilation in control group at 4 days, but then significantly higher at 7 days.

c. Patient outcomes: As mentioned in the validity section, they had a significant number of dropout patients for their data analysis. The significant clinical outcomes, however were appropriately assessed using an intention to treat analysis of all 146 patients entered into the study. These data are from the intention to treat analysis.

a) Significant decrease in ventilator days (11.0 +/- 1.1 treated vs. 16.3 +/- 1.9 controls, p=0.011), b) significant increase in ventilator-free days (17.6 treated vs. 12.7 controls, p=0.02), c) significant decrease in days on supplemental oxygen (15.8 ± 1.5 treated vs. 20.2 +/- 2.1 controls, p=0.053), d)significantly decreased length of stay in ICU (12.8 +/- 1.1 treated vs. 17.5 +/- 1.7 controls, p=0.016, e) significantly increased ICU-free days 16.0 treated vs. 11.0 controls, p=0.01)), f) significantly decreased new organ failures (8% vs. 28% c, p=0.015), and g) no significant difference in mortality ( 16% treated vs. 25% controlled, p=0.165).

d. Safety: No significant differences. Equal time to reach 75% of BEE, and tolerance to the feeding, no difference in serum triglycerides. Low incidence of gastrointestinal adverse events.

IV. Will the results help me in caring for my patients?

1. Are the likely treatment benefits worth the potential harms and costs? (Offer intervention on the basis of surrogate data only if patient's risk of the target outcome is high, patient places a high value on avoiding the target outcome, and if there are no satisfactory alternative therapies.)

Maybe yes. The addition of enteral formula that may be immunomodulatory to the treatment of systemic illnesses is an appealing therapy. If it can be given safely, then the risks of using it are very low. Most of the safety evaluations center around GI tolerance and serum lipid/triglyceride levels. Not much is known about other potential risks that these lipids may cause. Increased costs of this formula may be a significant.

This study satisfies the key criteria for a well done therapy article. Therefore the results of the paper should be considered when taking care of these patients. The question is whether all clinically important outcomes were considered. This therapy in ARDS patients appears to improve several surrogate end points, which may or may not be clinically important. BAL measures of inflammation, and P/F ratios make a lot of sense but have not been shown to be consistently correlated with mortality. Improvement of new organ failure seems the most significant correlate with mortality and was improved with this therapy. It remains to be seen what the eventual benefit of decreased ICU and ventilator days has on mortality in ARDS, but the obvious cost benefit is clear.

References

1. Fuhrman BP, et al. Futility of randomized, controlled ARDS trials--a new approach is needed. Crit Care Med. 1999;27:431-3. [citation]
2. Brochard L, et al Clinical trials in acute respiratory distress syndrome: what is ARDS? Crit Care Med. 1999;27:1657-8. [citation]
3. Davidson TA, et al. Reduced quality of life in survivors of acute respiratory distress syndrome compared with critically ill control patients. JAMA. 1999;281:354-60. [abstract]
4. Monchi M, et al. Early predictive factors of survival in the acute respiratory distress syndrome. A multivariate analysis. Am J Respir Crit Care Med. 1998;158:1076-81. [abstract]
5. Doyle RL, et al. Identification of patients with acute lung injury. Predictors of mortality. Am J Respir Crit Care Med. 1995;152:1818-24. abstract]
6. Jacobs S, et al. The Multiple Organ Dysfunction score as a descriptor of patient outcome in septic shock compared with two other scoring systems. Crit Care Med. 1999; 27: 741-745. [abstract]

Comments

Back to the EB Journal Club Index