QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Heyland, D. K. Search for Related Content PubMed PubMed Citation Articles by Heyland, D. K.

---

Supplement

In Search of the Magic Nutraceutical: Problems with Current Approaches¹

Department of Medicine, Queen’s University, Kingston, Ontario, Canada

	ABSTRACT

TOP ABSTRACT INTRODUCTION The use of surrogate... Underpowered studies The methodological quality of... Limited generalizability Believability of industry... CONCLUSIONS LITERATURE CITED

 
Over the last few decades, substrates with immune-modulating properties have been identified in all groups of micro- and macronutrients. Numerous experimental studies have focused on evaluating these substances, either alone or in combination. After hundreds of experiments, no clear, consistent signal exists that any of these agents result in significant treatment benefits in critically ill patients. The current approach to establishing the efficacy of nutritional interventions suffers from several limitations. First, the majority of studies focus on surrogate or substitute end points rather than clinically important end points. Second, the majority of clinical studies are small, and as such are underpowered to detect a significant treatment effect on clinically important end points. Third, the methodological quality of individual randomized trials varies. Methodological limitations, prevalent in nutrition studies, limit the strength of clinical inference that can be made from study results. High quality studies have been shown to differ significantly from low quality studies in their estimation of treatment effect. Fourth, the generalizability of single-site studies is limited. Finally, studies sponsored solely by industry are considered to be less believable than studies conducted under the auspices of peer-review agencies. Future evaluations must be done in the context of large, multicenter, well-designed, randomized trials focusing on clinically important end points that are sponsored from a variety of sources (including peer-reviewed agencies). Although such trials are costly, they are feasible and are much more likely to be believable and generalizable than the current approach.

KEY WORDS: • immunonutrition randomized clinical trials

	INTRODUCTION

TOP ABSTRACT INTRODUCTION The use of surrogate... Underpowered studies The methodological quality of... Limited generalizability Believability of industry... CONCLUSIONS LITERATURE CITED

 
Over the last few decades, substrates with immune-modulating properties have been identified in all groups of micro- and macronutrients. Numerous experimental studies have focused on evaluating these substances [glutamine, arginine, (n-3) fatty acids and others], either alone or in combination. For example, as outlined elsewhere in this publication, studies have measured the effect of glutamine on gastrointestinal mucosal atrophy, permeability and s-immunoglobulin A levels, nitrogen balance, plasma glutamine levels, immune function and cytokine levels. In the final analysis, what effect does glutamine have on the whole body? More importantly, on the critically ill patient?

To best answer these questions of efficacy, studies that compare the outcomes of two or more interventions are needed. A variety of methodologies can be used to compare the outcomes (Fig. 1 ). Each of these methodologies has its strengths, weaknesses and limitations but generally speaking, systematic reviews of randomized trials (and meta-analyses) and individual randomized trials tend to produce more robust results than cohort studies or case-control studies. The conclusion that more robust methods generate stronger inferences has led to the notion of levels of evidence and grades of recommendations (Sackett 1989 ). Strong clinical recommendations can be made (i.e., grade A recommendations) when supported by rigorous randomized trials or systematic reviews with a low chance of error (level I evidence). Moderately strong recommendation (grade B) can be made from randomized trials or systematic reviews with a high risk of error (level II evidence). Weaker recommendations (grade C) are based on less rigorous studies or randomized trials in different patient populations or randomized trials focusing on surrogate outcomes. Finally, no recommendations can be made from evidence that comes from nonrandomized studies of patient populations other than the one of interest, from animal studies or from studies based on biological rationale. A version of the relationship between levels of evidence and grades of recommendations is outlined in Table 1 .

View larger version (12K): [in this window] [in a new window]   Figure 1. Levels of evidence presented in hierarchical order from least potential for bias to most potential for bias. Abbreviation: RCT, randomized clinical trials.

	The use of surrogate outcomes

TOP ABSTRACT INTRODUCTION The use of surrogate... Underpowered studies The methodological quality of... Limited generalizability Believability of industry... CONCLUSIONS LITERATURE CITED

 
In the context of clinical trials, there are a variety of outcomes that could be measured. These range from nonclinically important to clinically important with surrogate or substitute end points in between (Fig. 2 ). The first problem with the current approach to establishing the efficacy of nutritional interventions is that the majority of studies focus on surrogate or substitute end points rather than clinically important end points. Surrogate end points are usually a laboratory measurement or physical sign used as a substitute for the clinically meaningful end point that directly measures how the patients feels, functions or survives. The expectation is that changes induced by a therapeutic intervention on a surrogate end point are expected to reflect changes in a clinically meaningful end point. With respect to the outcomes of RCT, because our objective is to improve the health outcomes of our patients, we should be concerned primarily with studies that evaluate the effect of health interventions on clinically important outcomes (outcomes that patients experience and are important to them). The problem is that these surrogate end points do not always correlate with clinically important health outcomes (Fleming and DeMets 1996 ). The "growth hormone" story in critically ill patients offers an illustrative example. Critical illness is associated with catabolism, protein turnover, negative nitrogen balance and prolonged weakness, resulting in prolonged ventilatory dependence and rehabilitation. Attempts to reverse these adverse effects of the catabolic response in critical illness have focused on the use of anabolic agents. Growth hormone has been shown to improve anabolism, nitrogen balance, wound healing and facilitate weaning from mechanical ventilation in several small studies. It was therefore considered as a therapeutic option in critically ill patients. However, in two large randomized trials in patients with prolonged critical illness, administration of growth hormone was associated with increased morbidity and mortality (Takala et al. 1999 ). Several other therapeutic interventions have been shown to have a significant effect on the surrogate outcome where later trials demonstrate that these interventions actually increase mortality rates [e.g., suppressing extraventricular premature beats (Pratt and Moye 1990 ), maximizing oxygen delivery in critically ill patients (Heyland et al. 1996 )].

View larger version (19K): [in this window] [in a new window]   Figure 2. Possible outcomes of clinical trials.

	Underpowered studies

TOP ABSTRACT INTRODUCTION The use of surrogate... Underpowered studies The methodological quality of... Limited generalizability Believability of industry... CONCLUSIONS LITERATURE CITED

 
The second problem in the current approach to establishing the efficacy of nutritional interventions is that the majority of clinical studies are small and as such, underpowered to detect a significant treatment effect on clinically important end points. This point is best illustrated by a meta-analysis of parenteral nutrition in critically ill patients (Heyland et al. 1998 ). There have been 26 randomized trials of surgical and critically ill patients examining the effect of parenteral nutrition on complication rates and mortality. These studies ranged in size from 18 to 395 patients with the majority of studies including <100 patients. The mortality event rate in these individual studies ranged from 0 to 41% with an overall average mortality rate of 8.9%. Individually, the majority of these studies were too small and underpowered to demonstrate a significant effect of parenteral nutrition on major complications or mortality. This is visually represented by the wide, imprecise confidence intervals (CI) around the point estimate of each individual study in the meta-analysis (Fig. 3 ). The advantage of a meta-analysis is that it provides a method of aggregating similar studies to come up with a more precise estimate of the treatment effect.

View larger version (33K): [in this window] [in a new window]   Figure 3. Risk ratios and associated 95% confidence intervals for the effect of total parenteral nutrition (TPN) on mortality in critically ill patients. [With permission from Heyland et al. (1998) .]

	The methodological quality of primary studies

TOP ABSTRACT INTRODUCTION The use of surrogate... Underpowered studies The methodological quality of... Limited generalizability Believability of industry... CONCLUSIONS LITERATURE CITED

View larger version (28K): [in this window] [in a new window]   Figure 4. Making inferences from randomized trials.

View larger version (20K): [in this window] [in a new window]   Figure 5. The methodological quality of trials of immunonutrition in critically ill patients. Abbreviation: CI, confidence interval.

Unblinded studies are susceptible to bias and weaken or invalidate inferences from clinical studies. For example, in a recent randomized, unblinded, multicenter trial of trauma patients, the efficacy of an "immune-enhancing" formula was compared with a standard enteral product (Moore et al. 1994 ). Study feedings were started within 24 h of admission to the Intensive Care Unit (ICU). On average, patients receiving the "immune-enhancing" feeding were on the ventilator for 1.9 ± 0.4 d vs. 5.3 ± 2.9 d for patients receiving the control feeding. It is not likely that a nutrition intervention initiated on day 1 in the ICU would result in patients coming off the ventilator 1 day later. Perhaps the baseline characteristics of these patients were somehow systematically different or they were managed differently compared with control patients. Unfortunately, the manuscript does not provide any details on other cointerventions, especially how patients were weaned from mechanical ventilators. The likelihood of such a bias weakens (if not invalidates) any inferences that can be made from this study. Admittedly, in some cases, blinding is not possible. In such situations, to minimize bias, investigators must make extra efforts to standardize patient care and/or use third-party (blinded) adjudicators to determine study outcomes.

Increasingly, studies of enteral nutrition are basing clinical recommendations on the results of one or more "compliance" analyses (Atkinson et al. 1998 , Bower et al. 1995 , Kenler et al. 1996 ). For example, Atkinson et al. (1998) reported a large, multicenter trial that compared an immune-enhancing formula with a standard formula in critically ill patients. Three hundred ninety-eight patients were entered into the study. On an intention-to-treat basis, 48% of the patients who received the immune-enhancing formula died compared with 44% in the control group (P > 0.05). There was no significant difference in length of stay. When the investigators restricted the data analysis to the 25% of the randomized patients who received a specific amount of enteral nutrition within the first 72 h, they found a significant reduction in duration of mechanical ventilation and length of ICU stay in patients receiving the immune-enhancing formula. However, patients receiving immunonutrition also tended to have an increased mortality and to die earlier (which may explain why length of stay was reduced!). The authors concluded, "... those patients in whom it was possible to achieve early enteral nutrition ... had a significant reduction in the morbidity of their critical illness." An editorial accompanying the publication of this study (Zaloga 1998 ) stated, "... the analysis of subgroups who received critical amounts of formula is valid. Intention-to-treat analysis is less valid since it would also fail to demonstrate the beneficial effects of drugs in patients who fail to receive adequate amounts."

This notion is inconsistent with current teaching in clinical epidemiology (Guyatt et al. 1993 ). An intention-to-treat analysis requires that all patients originally randomized in the trial be accounted for in the group to which they were randomized and is generally considered to be the most valid estimate of treatment effect. The goal of randomization is to equally distribute known and unknown prognostic variables between study groups so that observed differences can be attributed to the study intervention and not differences in patient populations. When randomized patients are dropped or excluded from the analysis, the distribution of these prognostic variables may be disturbed and the study results biased. Because they may include patients who do not receive the intervention, intention-to-treat analyses can be considered to be less sensitive in detecting a treatment effect.

Another approach to analyzing the data that represents a compromise between validity and sensitivity is known as an "efficacy" analysis. In an efficacy analysis, the investigator states a priori that certain groups of patients will be omitted from the analysis on the basis of scientific rationale. For example, patients who were not "truly eligible" or patients who never received the feedings might be excluded. Compared with the intention-to-treat analysis, this approach is more sensitive to detecting a treatment effect but is more likely to be biased. The burden rests with the investigators to convince the scientific community that the analysis has not been biased.

A third approach to analyzing data is to group patients on the basis of events that occur after randomization, such as the amount of enteral feedings received. Analyses classified by variables measured after baseline (such as compliance or tolerance) are more likely to mislead than to inform (Oxman and Guyatt 1992 , Yusuf et al. 1991 ). There may be an interaction between compliance or tolerance to enteral feedings and the study intervention, resulting in a bias that cannot be compensated. For example, most immune-enhancing formulas contain arginine. Arginine supplementation increases nitric oxide production. Nitric oxide plays a role in gastric emptying (Zaloga and Marik 2000 ). Therefore, arginine supplementation may decrease gastric emptying, resulting in decreased tolerance of enteral feedings in the experimental group. Patients in the experimental group may not tolerate enteral feeding for reasons that are different from those of patients in the control group, resulting in different patient populations in the "compliance" analysis. For this reason, it is likely that these analyses are biased. It is recommended that no study subjects be withdrawn from the analysis because of compliance reasons (Friedman et al. 1985 ).

Returning to our clinical example (Atkinson et al. 1998 ), the investigators stated, a priori, that only patients who tolerated >2.5 L within 72 h would be included in the analysis (just because the investigators state a priori their decision rule does not eliminate the risk of bias). If only 25% of randomized patients are compliant with this amount, it raises questions concerning the utility of the intervention. Furthermore, patients purported to benefit cannot be identified a priori. Therefore, if this benefit is to be gained, all of the patients would have to be treated. However, we already know that when all of them were treated, no advantage ensued. Therefore, there must be some subgroup in whom the treatment was disadvantageous. Whatever gain is achieved in one subgroup would be lost in another. The good would be negated by harm.

	Limited generalizability

TOP ABSTRACT INTRODUCTION The use of surrogate... Underpowered studies The methodological quality of... Limited generalizability Believability of industry... CONCLUSIONS LITERATURE CITED

 
The fourth problem plaguing the current approach is the limited generalizability of single-site studies. To ensure that the clinical and economic results of randomized trials are applicable to patients in other settings, both the patients in the study and their management have to be similar to what is seen in other settings. Important differences in case-mix, practice patterns and health care systems will limit the generalizability of results. Clinical trials emanating from single centers are less likely to be generalizable to broader settings. Because multicenter studies include patients and management styles with greater diversity, their results are more likely to be applicable to a broader range of settings. Of the 26 randomized trials included in the meta-analyses of parenteral nutrition (Heyland et al. 1998 ) and the 22 studies of immunonutrition (Heyland et al. 2000 ), only 3 of 26 (12%) and 8 of 22 (36%), respectively, were multicenter studies. If we are to improve the generalizability of randomized controlled trials of nutrition support, these studies must involve more than one center.

	Believability of industry-sponsored research

TOP ABSTRACT INTRODUCTION The use of surrogate... Underpowered studies The methodological quality of... Limited generalizability Believability of industry... CONCLUSIONS LITERATURE CITED

 
The final problem with the current approach to conducting randomized trials is that they depend heavily on industry sponsorship. Companies manufacturing or selling nutrition-related products or devices for profit play an enormous role in research and education. Many continuing medical education events at local and international levels and peer-reviewed journals would not be possible without the generous support from industry partners. Although published statistics are lacking, industry partners finance the majority of medical and nutrition research. In the aforementioned meta-analysis of immunonutrition, only 4 of 22 studies(18%) mentioned support from peer-reviewed agencies (Heyland et al. 2000 ).

Although many physicians implicitly mistrust the results of studies sponsored by industry, there is a growing body of empirical evidence to support this notion. It is clear that interactions between physicians and the pharmaceutical industry influence prescribing and professional behavior (Wazana 2000 ). Considerable difficulties are encountered in attempts to access original data for the purposes of conducting a meta-analyses when those data reside with pharmaceutical companies (Johansen and Gotzsche 1999 ). Other studies suggest that physicians with financial ties to manufacturers were much less likely to criticize the safety or efficacy of these agents (Barnes and Bero 1998 , Rochon et al. 1994 , Stelfox et al. 1998 ). Similarly, in a study of publications of clinical trials pertinent to the practice of general internal medicine, there was a significant association between positive results and funding from pharmaceutical manufacturers (Davidson 1986 ). Negative results sometimes are not even published while product continues to be sold (Ross Products 1996 ). Finally, in a recent review of economic evaluations, Friedberg et al. (1999) found that pharmaceutical company sponsorship of economic analyses was associated with a reduced likelihood of reporting unfavorable results. Collectively, these findings raise serious concerns about the believability of industry sponsored trials.

These concerns can have an adverse effect on physician practice in the following two ways: 1) physicians can adopt new practices based on misleading information (false positive) from industry trials; or 2) because of their inherent disbelief of industry-sponsored research, some physicians may not adopt new technology that is shown (true positive) to improve patient outcomes (and economic outcomes). Alternative sources of funding, either alone or in combination with industry support would likely enhance the believability of randomized trials, although empirical evidence for this notion is lacking.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION The use of surrogate... Underpowered studies The methodological quality of... Limited generalizability Believability of industry... CONCLUSIONS LITERATURE CITED

 
The results of randomized trails are intended to inform physicians, patients and decision makers regarding the relative merits of nutritional (medical) interventions. The current approach to determining the efficacy of these interventions is characterized by small, single-center, randomized trials of varying methodological quality and sponsored primarily by industry sources. Despite considerable resources spent conducting these trials of various immune-enhancing nutrients, including glutamine, due to the deficiencies of this approach, there are still no clear answers concerning the relative efficacy of these substances in most patient populations. If we are to make advances in our knowledge about the role of glutamine (or any other nutritional intervention), we must conduct large, multicenter, well-designed, randomized trials that are supported from more sources than industry partners and that evaluate the effect of nutrient supplementation on clinically important end points. Although such trials are costly, they are feasible and are much more likely to yield definitive, valid results and be generalizable than the current clinical trials.

	FOOTNOTES

 
¹ Presented at the International Symposium on Glutamine, October 2–3, 2000, Sonesta Beach, Bermuda. The symposium was sponsored by Ajinomoto USA, Incorporated. The proceedings are published as a supplement to The Journal of Nutrition. Editors for the symposium publication were Douglas W. Wilmore, the Department of Surgery, Brigham and Women’s Hospital, Harvard Medical School and John L. Rombeau, the Department of Surgery, the University of Pennsylvania School of Medicine.

² Abbreviations used: CI, confidence interval; ICU, Intensive Care Unit; RCT, randomized clinical trials; RR, risk ratio.

	LITERATURE CITED

TOP ABSTRACT INTRODUCTION The use of surrogate... Underpowered studies The methodological quality of... Limited generalizability Believability of industry... CONCLUSIONS LITERATURE CITED

 

1. Atkinson S., Sieffert E. & Bihari D. (1998) A prospective, randomized double-blind, controlled trial of enteral immunonutrition in the critically ill. Crit. Care Med. 26:1164-1172.[Medline]

2. Barnes D. E. & Bero L. A. (1998) Why review articles on the health effects of passive smoking reach different conclusions. J. Am. Med. Assoc. 279:1566-1570.[Abstract/Free Full Text]

3. Bower R. H., Cerra F. B., Bershadsky B., Licari J. J., Hoyt D. R., Jensen G. O., Van Buren C. T., Rothkopf M. P., Daly J. O. & Adelsberg B. R. (1995) Early enteral administration of a formula (Impact) supplemented with arginine, nucleotides, and fish oil in intensive care unit patients: results of a multicenter, prospective, randomized, clinical trial. Crit. Care Med. 23:436-449.[Medline]

4. Chalmers T. C., Celano P., Sacks H. S. & Smith H. (1988) Bias in treatment assignment in controlled clinical trials. N. Engl. J. Med. 309:1358-1361.[Abstract]

5. Davidson R. A. (1986) Source of funding and outcome of clinical trials. J. Gen. Intern. Med. 1:155-158.[Medline]

6. Fleming T. R. & DeMets D. L. (1996) Surrogate end points in clinical trials: are we being misled?. Ann. Intern. Med 125:605-613.[Abstract/Free Full Text]

7. Friedberg M., Saffran B., Stinson T. J., Nelson W. & Bennett C. L. (1999) Evaluation of conflict of interest in economic analyses of new drugs used in oncology. J. Am. Med. Assoc. 282:1453-1457.[Abstract/Free Full Text]

8. Friedman L. M., Furberg C. D. & DeMets D. L. (1985) Fundamentals of Clinical Trials 2nd ed. 1985:241-249 PSG Publishing Co Littleton, MA. .

9. for the Evidence-Based Medicine Working GroupGuyatt G. H., Sackett D. L. & Cook D. J. (1993) Users’ guides to the medical literature. II. How to use an article about therapy or prevention. A. Are the results of the study valid?. J. Am. Med. Assoc. 270:2598-2601.[Free Full Text]

10. for the Evidence-Based Medicine Working GroupGuyatt G. H., Sackett D. L. & Cook D. J. (1994) Users’ guides to the medical literature. II. How to use an article about therapy or prevention. B. What were the results and will they help me in caring for my patients?. J. Am. Med. Assoc. 271:59-63.[Free Full Text]

11. Heyland D. K., Cook D. J. & Guyatt G. H. (1993) Enteral nutrition in the critically ill patient: a review. Intensive Care Med 19:435-442.[Medline]

12. Heyland D. K., Cook D., King D., Kernerman P. & Bruin-Buisson C. (1996) Maximizing oxygen delivery in critically ill patients: a methodologic appraisal of the evidence. Crit. Care Med. 24:517-524.[Medline]

13. Heyland D. K., MacDonald S., Keefe L. & Drover J. W. (1998) Total parenteral nutrition in the critically ill patient: a meta-analysis. J. Am. Med. Assoc. 280:2013-2019.[Abstract/Free Full Text]

14. Heyland D. K., Novak F., Drover J., Jain M. & Suchner U. (2000) Should immunonutrition become routine in critically ill patients: A systematic review of the evidence. J. Am. Med. Assoc. 2001 (in press).

15. Johansen H. K. & Gotzsche P. C. (1999) Problems in the design and reporting of trials of antifungal agents encountered during meta-analysis. J. Am. Med. Assoc. 282:1752-1759.[Abstract/Free Full Text]

16. Kenler A. S., Swail W. S., Driscoll D. F., DeMichele S. J., Daley B., Babineau T. J. & Peterson M. B. (1996) Early enteral feeding in postsurgical cancer patients: fish oil structured lipid-based polymeric formula versus a standard polymeric formula. Surgery 223:316-333.

17. Moore F. A., Moore E. E., Kudsk K. A., Brown R. O., Bower R. H., Koruda M. J., Baker C. C. & Barbul A. (1994) Clinical benefits of an immune-enhancing diet for early postinjury enteral feeding. J. Trauma 37:607-615.[Medline]

18. Oxman A. D. & Guyatt G. H. (1992) Apples, oranges and fish: a consumer’s guide to subgroup analyses. Ann. Intern. Med. 116:78-84.

19. Pratt C. M. & Moye L. A. (1990) The cardiac arrhythmia suppression trial: background, interim results and implications. Am. J. Cardiol. 65:20B-29B.[Medline]

20. Rochon P. A., Gurwitz J. H., Simms R. W., Fortin P. R., Felson D. T., Minaker K. L. & Chalmers T. C. (1994) A study of manufacturer-supported trials of nonsteroidal anti-inflammatory drugs in the treatment of arthritis. Arch. Intern. Med. 154:157-163.[Abstract]

21. Ross Products Division of Abbott Laboratories (1996) Comparison of Option One and a Polymeric Enteral Feeding: Effect on Length of Stay and Clinical and Immune Parameters: Study Protocol 1996 Ross Products Columbus, OH. .

22. Sackett D. L. (1989) Rules of evidence and clinical recommendations on the use of antithrombotic agents. Chest 95:2S-4S.[Free Full Text]

23. Sacks H. S., Chalmers T. C. & Smith H., Jr (1983) Randomized versus historical assignment in controlled trials. N. Engl. J. Med. 309:1353-1361.[Abstract]

24. Schultz K. F., Chalmers I., Hayes R. J. & Altman D. G. (1995) Empirical evidence of bias: dimensions of methodological quality associated with estimates of treatment effects in controlled trials. J. Am. Med. Assoc. 273:408-412.[Abstract]

25. Stelfox H. T., Chua G., O’Rourke K. & Detsky A. (1998) Conflict of interest in the debate over calcium-channel antagonists. N. Engl. J. Med. 338:101-106.[Abstract/Free Full Text]

26. Takala J., Ruokonen E., Webster N. R., Nielsen M. S., Zanstra D. F., Vundelinckx G. & Hinds C. J. (1999) Increased mortality associated with growth hormone treatment in critically ill patients. N. Engl. J. Med. 341:785-792.[Abstract/Free Full Text]

27. Wazana A. (2000) Physicians and the pharmaceutical industry. J. Am. Med. Assoc. 283:373-380.[Abstract/Free Full Text]

28. Yusuf S., Wittes J., Probstfiel J. & Tyroler H. A. (1991) Analysis and interpretation of treatment effects in subgroups of patients in randomized clinical trials. J. Am. Med. Assoc. 266:93-98.[Abstract]

29. Zaloga G. P. (1998) Immune-enhancing enteral diets: where is the beef?. Crit. Care Med. 26:1143-1145.[Medline]

30. Zaloga G. P. & Marik P. (2000) Promotility agents in the intensive care unit. Crit. Care Med. 28:2657-2659.[Medline]

This article has been cited by other articles:

				C. Duggan, J. Gannon, and W A. Walker Protective nutrients and functional foods for the gastrointestinal tract Am. J. Clinical Nutrition, May 1, 2002; 75(5): 789 - 808. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Heyland, D. K. Search for Related Content PubMed PubMed Citation Articles by Heyland, D. K.