1. Top of page
  2. Abstract
  10. References

Non-steroidal anti-inflammatory drugs induce damage throughout the entire gastrointestinal tract. Administration of site-specific permeability probes is a non-invasive technique for assessing the functional integrity of the gastrointestinal mucosa. A systematic search for NSAID-induced permeability studies using MEDLINE and EMBASE, and an analysis of the literature on NSAID-induced gastrointestinal permeability, were carried out. The advantages and disadvantages of the various probes and study protocols are discussed.

Identification of the underlying mechanisms of regulatory control of the epithelial tight junction is still needed. A greater appreciation of the pharmacokinetics and distribution of NSAIDs, coupled with gastrointestinal permeability studies, may help delineate the pathogenesis of NSAID-induced gastrointestinal toxicity. Non-invasive tests of gastric, intestinal and colonic permeability have shown promise in both basic research and in clinical practice. While such tests could not replace endoscopy, they may represent clinically useful techniques for identifying patients who would benefit from endoscopy, to assess the response to treatment, and perhaps to predict the clinical course of disease.


  1. Top of page
  2. Abstract
  10. References

The anti-inflammatory, analgesic, anti-pyretic and anti-thrombotic properties of non-steroidal anti-inflammatory drugs (NSAIDs) make this therapeutic class of particular utility in the management of chronic arthropathies, in the treatment of pain, in the alleviation of fever, and for the prevention of myocardial infarction and stroke. In addition, recent evidence suggests possible therapeutic benefits of NSAIDs in the prevention of colorectal cancer and in the treatment of Alzheimer’s disease.1, 2

There are currently more than 35 NSAIDs available for clinical use world-wide.3 Previous estimates of their use include >100 million prescriptions written throughout the world, at a cost of well over $5 billion.4 All NSAIDs, when given in equivalent therapeutic doses, demonstrate comparable efficacy.5, 6 Given this apparently equivalent efficacy, the relative safety profile of individual NSAIDs is becoming a principal criterion for therapeutic selection. The clinical utility of NSAIDs is determined as a compromise between therapeutic efficacy and acceptable side-effects. However, the ‘safety’ of NSAIDs is promotionally driven, as it is a main determinant of the success of an NSAID’s marketability. Unfortunately, there does not appear to be a consensus as to what constitutes a ‘safer’ NSAID. There is no precise definition of ‘safer’ in terms of symptomatic gastrointestinal (GI) side-effects, short-term endoscopy studies, 6-month endoscopy studies or clinically significant GI complications. More rigorous scientific substantiation of claims of reduced NSAID toxicity is required in the development of GI-safe NSAIDs.

The GI side-effects of NSAIDs were recognized more than a century ago.7 The clinical use of NSAIDs has been associated with numerous side-effects, the most important in frequency and clinical impact being GI disturbance.8 NSAID GI pathology accounts for more than 70 000 hospitalizations and 7000 deaths annually in the USA.9 Adverse effects in the GI tract have contributed to the termination of clinical development and also the withdrawal from the market of several NSAIDs.8 Epidemiological studies indicate significant differences in the incidence of various NSAID-induced adverse effects, including those occurring in the upper GI tract.6, 9[Committee on the Safety of Medicines. ]–11

Numerous articles have been published examining the gastric and duodenal damage caused by NSAIDs. However, there is often a lack of correlation between gastric symptoms and gastroscopic evidence of ulcers, with as many as one-third of patients being completely asymptomatic.12 Moreover, amongst patients with upper GI lesions and blood loss, healing of the lesions is not always accompanied by an improvement in anaemia, which may suggest the occurrence of NSAID-induced afflictions in more distal sites of the GI tract.13 There is a growing body of evidence that more distal GI damage may be widespread and of more serious consequence than previously thought.

The distal intestinal disturbances caused by NSAIDs have recently received closer attention.14[15]–16 It has been suggested that the prevalence of lower GI side-effects may be higher than that detected in the upper GI tract and may be of major clinical significance.17 In a 1985 epidemiological study, the expected incidence of lower bowel perforations and bleedings was determined to be 10 and 7 per 100 000, respectively.18 Although the increased incidence of small intestinal ulceration in patients prescribed NSAIDs is less commonly clinically appreciated than those in the stomach or duodenum, the serious clinical manifestations of the ulcers (bleeding and perforation) may be life-threatening.19 It has also been reported that 41% of rheumatoid arthritis (RA) patients taking NSAIDs with iron deficiency anaemia and undiagnosed GI blood loss had evidence of small intestinal lesions, erosions and ulcers iatrogenically attributed to NSAIDs upon small bowel enteroscopy.13 Blood loss from the lower GI tract may result in significant morbidity contributing to the anaemia of RA patients taking NSAIDs. Vitamin B12 and bile acid absorption may also be impaired, contributing to anaemia and increasing morbidity.13, 20 Some studies have demonstrated that up to 70% of patients taking NSAIDs chronically develop intestinal inflammation associated with blood and protein loss; on discontinuation of NSAIDs this intestinal inflammation may persist for up to 16 months.21[22]–23 Further studies have shown that long-term NSAID treatment leads to enhanced migration of 111indium-labelled leucocytes, predominantly to the mid-small intestine, which suggests mucosal inflammation of the small bowel. This, together with evidence of increased faecal 111indium excretion, provides further evidence that NSAIDs cause intestinal inflammation in a substantial number of patients chronically receiving these drugs.20, 24, 25 Considering the extent of world-wide NSAID use, the clinical manifestations in the distal GI tract (bleeding, perforation and ulceration) undoubtedly contribute to significant morbidity in many patients.

Because NSAIDs have been linked to the development of serious GI side-effects, numerous strategies have been employed to reduce this mucosal damage before it occurs. Various approaches have been taken to this problem, including the development of prodrugs, once daily dosing (long t1/2 NSAIDs), and enteric-coated and modified-release formulations. An alternative approach has been the concomitant treatment with protective substances to circumvent NSAID-induced GI side-effects. Preventative measures evaluated to-date have utilized a wide variety of pharmacological approaches, including antisecretory agents (H2-receptor antagonists, proton pump blockers, anticholinergic agents and antacids); as well as attempts to increase mucosal defence (sucralfate, and prostaglandin analogues). However, none of these approaches has solved the problem of NSAID-induced GI damage.

In recent years, there have emerged three approaches to the development of new ‘GI-safe’ NSAIDs. Intensive efforts are now being made to develop selective inhibitors of cyclooxygenase-2 (COX2), assuming that these agents will inhibit this isoform when it is induced at sites of inflammation, but will not inhibit prostaglandin synthesis in other tissues such as the stomach, where cyclooxygenase-1 (COX1) is constitutively expressed.26 Another novel strategy to reduce the GI ulcerogenicity of NSAIDs that has recently been described is the incorporation of a nitric oxide (NO) generating moiety into the NSAID molecule. NO may counteract the detrimental effects of COX suppression such as maintaining blood flow, and prevent leucocyte adherence such that mucosal damage does not occur.27 Finally, the pre-association of NSAIDs with zwitterionic phospholipids (DPPC–NSAIDs) may reduce the ability of NSAIDs to associate with phospholipids in the mucus gel, and therefore may reduce ulcerogenicity.28

The main clinical goal still remains to solve the problem of NSAID-induced GI toxicity, with its resultant morbidity and mortality, while obtaining optimal therapeutic effect. Although there are several alternatives to traditional NSAIDs emerging (i.e. COX2 inhibitors, NO–NSAIDs, DPPC–NSAIDs) that hold such promise, any new therapeutic agent, no matter what the rationale behind its introduction, must be assumed to induce GI toxicity until clinically proven otherwise.


  1. Top of page
  2. Abstract
  10. References

A major problem in the care of patients on long-term NSAID therapy is the diagnosis of GI complications induced by these agents. Gastroduodenal endoscopy has become the gold standard method for assessing NSAID-induced GI damage, because it has been generally believed that the side-effects of NSAIDs are usually confined to the gastroduodenal mucosa. Endoscopy for the diagnosis of upper gastrointestinal disease is widely available, precise, sensitive and easy to perform. However, endoscopy for diagnosis of upper GI damage induced by NSAIDs is unsuitable as a routine screening test as it is time-consuming, expensive and may not be available in all centres.

The majority of patients with endoscopically-determined NSAID-induced gastric damage are asymptomatic.29, 30 Clinical diagnosis of NSAID-induced gastroduodenal abnormalities is also associated with difficulties such as the determination of the site of affliction and difficult distinctions between degrees of damage. There is little consistency in the published literature on the type of injury being assessed. The gastroduodenal lesions associated with NSAIDs may be described as oedema, erythema, mucosal haemorrhage, erosions or ulcers. The distinction between erosions and ulcers is not clear and can vary among clinicians.31 Investigators often give numerical values to each type of lesion and total them up to obtain a score for each patient. The Lanza scale has been demonstrated to be relatively effective, with few inter- and intra-observer differences. However, this scoring system is also subjective and therefore susceptible to error and inconsistencies in interpretation.32 For example, in one study, an ulcer was defined as a break in the mucosa greater than 3 mm in diameter.33 This ‘ulcer’ could easily be described as an erosion by others. Such damage is later described by the same investigator as ‘trivial lesions’ and a re-definition of an ulcer was suggested to be damage greater than 5 mm or even 1 cm in diameter.34

Endoscopic detection as a surface view cannot give any information as to whether the lesion is actually a true ulcer that penetrates through to the muscularis mucosae, unless biopsies are also taken. It may be argued that an important ulcer is one which produces symptoms and that can be improved with treatment or on discontinuing NSAIDs, or that leads to a serious complication (bleeding, perforation or stricture). Detection upon endoscopy is usually only demonstrated after a complication (obstruction, perforation or haemorrhage) becomes clinically apparent. It has also been suggested that a distinction be made between ‘endoscopic’ ulcers, seen during routine endoscopic surveillance, and ‘clinically relevant’ ulcers, seen during investigation of complications, which may provide an approximation to the problem.31

It is still essential that a means of detecting NSAID-induced lesions exists before clinically significant sequelae and hospitalization is required. Since the early 1980s, substantial efforts have been made to develop non-invasive methods of detecting GI abnormalities.35, 36 As the intercellular junctions of GI epithelial cells appear to be particularly susceptible to a variety of noxious agents, they may be the first organelle to suffer when the energy production of the enterocyte is compromised. This disruption of intercellular integrity allows permeation of macromolecules into the GI mucosa. The degree of GI penetration by passively absorbed water-soluble molecules is referred to as permeability. Tests of GI permeability are designed to assess the functional integrity of the intestinal barrier. This is accomplished non-invasively by measurement of urinary recovery of a variety of orally administered probes. Methods based upon measurement of GI permeability have been found to be valuable research tools and have clinical utility in measuring gastroduodenal, intestinal and colonic damage induced by NSAIDs in numerous clinical studies ( Table 123 ).

Table 1.  . NSAID-induced 51Cr-EDTA intestinal permeability studies Thumbnail image of
Table 2. Table 1.(Cont)Thumbnail image of
Table 3. Table 1.(Cont)Thumbnail image of

The value of gastroscopy is that it can identify patients who are at risk of serious complications and therefore any non-invasive screening tests must also be able to provide this information in order to have clinical utility. It is still important that a means of detecting NSAID-induced lesions exists before clinically significant sequelae and hospitalization are required. Gastrointestinal permeability tests could be used to sequentially follow patients at risk of GI disease, such as those receiving NSAIDs. Through the clinical use of permeability tests it could also be therapeutically possible to intervene before clinically detectable GI disease becomes evident or reactivates.

With the development of GI-safe NSAIDs, the question may arise as to whether there will be a need for a screening test for NSAIDs in the future. Nevertheless, there may be other clinical applications of gastric permeability tests for other xenobiotics such as aminobisphosphonates and other diseases that afflict the GI tract, including lymphocytic gastritis, malaria, portal hypertensive gastropathy and gastric cancer.


  1. Top of page
  2. Abstract
  10. References


Small bowel enteroscopy has been successfully used to detect small intestinal lesions, erosions and ulcers attributed to NSAIDs.13 However, the clinical use of enteroscopy is currently not available in many centres.

Intestinal permeability tests in vivo have been assessed by a number of analytical techniques. The three most commonly employed markers are the urinary excretion following oral ingestion of carbohydrates (i.e. lactulose, cellobiose and mannitol), ethylene glycol polymers [i.e. polyethylene glycol (PEG)], and non-degradable radionuclide probes such as 51Cr-EDTA.37[38]–39

The movement of molecules across biological membranes may occur via (i) simple diffusion or (ii) specific transport mechanisms. The specific transport mechanisms include those that are carrier mediated such as (a) facilitated diffusion, (b) exchange diffusion (countertransport), (c) active transport and (d) pinocytosis.40

The permeability of the intestine is highly regulated, reflecting at least three distinct unmediated permeation pathways of mucosal diffusion: (i) the intercellular junction between adjacent enterocytes, (ii) aqueous pores in the enterocyte brush border membrane and (iii) the lipid-soluble rich hydrophobic pathway in the brush border. In vivo permeability tests assess the functional integrity of the intestinal barrier.

The requirements of an ideal passive permeability marker include: it should be non-toxic, absorbed entirely by passive diffusion, not modified or metabolized by enzymes, not found in the diet, not produced endogenously, cleared from the body rapidly and completely, hydrophilic, limited to the extracellular compartment, non-immunogenic, and easily and rapidly measurable in biological fluids with both high precision and accuracy.41[42]–43 All of the current intestinal permeability markers have advantages and disadvantages and none possess all the criteria of an ideal marker ( 4 Table 2). Other than their non-invasive nature, a distinct advantage of permeability tests is that they reflect the functional state of a major area of the intestinal mucosa, whereas morphological analysis may suffer due to sampling error—particularly if the GI abnormality is randomly distributed.

Table 4. Table 2. Some ideal requirements for an intestinal permeability markerThumbnail image of

Polyethyleneglycol (PEG)

Current interest in the measurement of intestinal permeability originated from the initial work undertaken with PEG.41 PEG is a viscous mixture of liquid polymers; the number corresponds to its average molecular weight. PEG 400 has recently been extensively employed as a permeability probe.44, 45

After an overnight fast, PEG (1–5 g in 50–100 mL of water) is ingested and urine is collected, usually from 0 to 6 h.46, 47 However, it has become evident that PEG is an unsatisfactory probe for measuring intestinal permeability.48 Urinary recovery of PEG polymers is low and variable following intravenous instillation in humans, indicating loss to peripheral body compartments.37 PEG is relatively lipophilic and this is reflected in extensive permeation through the GI mucosal membranes ( Figure 1). PEG is therefore a relatively insensitive marker for detecting small increases in mucosal permeability though the intercellular pathway, because the background permeation is extremely high and small changes in permeability are not easily detectable.41, 49 Also, the use of PEG to assess intestinal permeability in a variety of disorders have given results in the vast majority of studies that are contradictory to the differential urinary excretion of di/monosaccharides and are difficult to interpret.44, 45, 50, 51 Some studies of PEG 400 intestinal permeability are conflicting, because both increased and decreased permeability have been found.52, 53 Furthermore, laborious specimen preparation, an unpleasant taste, inter-batch variation, and demanding analytical techniques, also limit the utility of PEG as an intestinal permeability marker.


Figure 1. 51. Cr-EDTA and disaccharides pass exclusively through the intercellular junctions between enterocytes. Monosaccharide permeation occurs mainly through aqueous pores whereas PEG permeation is primarily determined by the lipophilic character of the brush border membrane.

Download figure to PowerPoint


A number of carbohydrates have been used in studies of permeability, including the hexoses (i.e. L-rhamnose), sugar alcohols (i.e. D-mannitol), and the disaccharides (e.g. lactulose, cellobiose). D-xylose has also been used in permeability studies but as it can be transported into the enterocyte by a carrier-mediated process, it is not an indicator of passive permeability. Xylose absorption also appears to be reduced after both indomethacin treatment in non-arthritic subjects and in indomethacin-treated patients with rheumatoid arthritis.54, 55

Most of the currently utilized carbohydrate permeability studies employ differential sugar absorption tests, in which two sugars (usually a mono- and a disaccharide) are given concomitantly and urinary recovery of each is determined. In differential sugar tests, patients ingest a mixture of hyperosmolar sugars (1330 mmol/kg of water) following an overnight fast, and urine is collected for 5 h.39, 56 The monosaccharide shows the degree of permeation through aqueous pores and the disaccharide reflects the permeation of the intercellular pathway. The ratio of the disaccharide to the monosaccharide gives an index of relative function of the permeability pathways.43

Travis & Menzies36 have demonstrated that permeation of lactulose and oligosaccharides is temporarily markedly increased in normal individuals when test dose osmolarity is increased beyond 1500 mOsm/L. This effect is evident with most readily absorbed solutes, but these changes vary in magnitude with each osmotic filler and there is also considerable individual variability in this effect.36 As a consequence there is practical importance in controlling the osmolarity of test solutions in intestinal permeability studies. The majority of investigators now use physiological iso-osmolar tests rather than osmotic fillers. In order for results to be compared directly between groups, standardization of test dose composition is essential in permeability studies.

The use of carbohydrate probes also has some inherent clinical problems. The osmotic fillers may affect mucosal permeability, because hypertonic lactulose test solutions have been shown to decrease permeability.57 Some carbohydrates are metabolized by intestinal flora and the brush border membrane can metabolize cellobiose, which will lead to an underestimation of permeability. In addition, rhamnose is incompletely excreted after intravenous instillation, which may also underestimate its intestinal permeability. Lactulose and mannitol may be present in some foodstuffs, and some patients may excrete minute quantities of endogenously produced mannitol, which may underestimate intestinal permeability.46, 58 Furthermore, sugar loads can cause abdominal distention, diarrhoea and flatulence in some patients.58 Additionally, quantitative analysis of carbohydrates generally requires specific chemical analysis or lengthy and time-consuming extraction and chromatographic procedures, which are often prohibitive for routine screening.

Studies using carbohydrate markers have produced apparently contradictory results. These misunderstandings have arisen because of the lack of insight into the relative sensitivity of the different methods used to assess intestinal permeability. For example, some investigators have reported that permeability measurements for the disaccharide, lactulose, agree closely (r2 = 0.82–0.98) with the permeability measurements obtained with 51Cr-EDTA,46, 59 while other investigators have noted that NSAID-induced permeability as measured by 51Cr-EDTA excretion did not result in a corresponding increase in mono- or disaccharide permeation.56, 60, 61 Subsequently, several reports39, 62 using urinary sugar excretion have confirmed the initial results of Bjarnason’s group17, 33, 63[64][65][66][67][68]–69 and have demonstrated abnormalities of intestinal permeability in RA patients on NSAIDs.

In addition to the theoretical requirements of an ideal permeability probe ( Table 12 ), the use of excretory ratios of markers (i.e. lactulose/mannitol; 51Cr-EDTA/mannitol; 51Cr-EDTA/ L-rhamnose; cellobiose/ L-rhamnose) to detect an alteration in the intestinal barrier may have advantages over a single marker by minimizing some of the extramural factors such as intestinal motility changes. Such extramural factors would affect both markers identically so that their urinary excretion ratios are unchanged, resulting in a decrease in the variability and an increase in the reliability of the intestinal permeability data. The value of the urine excretion ratio specifically reflects alterations in intestinal permeability and is largely unaffected by pre-mucosal and post-mucosal determinants of the overall permeability of the intestine, because each probe acts as an internal standard for the other.


51Cr-EDTA was initially used as a marker of glomerular filtration and subsequently as a screening test for coeliac disease.68, 7051Cr-EDTA incorporates the analytical advantages of a γ-ray emitting isotope in a water-soluble, highly stable, chelated compound with physicochemical inertness.69 The compound is stable and has a half-life of 1 month so that aliquots of the radioactive material can be easily stored.

After an overnight fast, subjects drink a test solution containing 100 μCi of 51Cr-EDTA in 10–100 mL of distilled water, followed by 300 mL of water. Subjects are required to fast for an additional 2 h, after which they are allowed food and fluid. Adequate sensitivity of the 51Cr-EDTA test seems to increase with a longer collection period of up to 24 h.38 Alcohol and spicy foods are prohibited for 3 days before and throughout each study. Results are expressed as a percentage of the orally administered test dose excreted in the urine during this time interval.

The permeation of 51Cr-EDTA has been shown to be relatively specific to the small intestine, given that a comparison of peroral and intraduodenal instillation showed no significant differences in the extent of urinary excretion in both animal and clinical studies.70[71]–72 However, more recent studies have suggested that there may be some degree of colonic permeation.61, 73 Nevertheless, it has been suggested that the percentage of colonic permeation is low because 51Cr-EDTA is incorporated into faeces, limiting the availability for colonic permeation.33 However, if there is significant epithelial damage to the colonic epithelium (i.e. in colitis) 51Cr-EDTA permeation may increase. The 51Cr-EDTA test has been shown to be reproducible and the safety, simplicity and accuracy of the procedure meet all the requirements for a permeability test for NSAID-induced intestinal permeability studies.

The 51Cr-EDTA test has received criticism for high inter-individual variability and low sensitivity.70, 74, 75 However, these studies have used small sample numbers and the reproducibility appears to be increased with a urine collection over 24 h. Alcohol may profoundly affect the results of this permeability test, but this is reversible upon abstinence.76 One may also question the theoretical possibility that membrane integrity could be affected by EDTA itself. However, experiments on animal tissue suggest that the concentrations required would need to be 1000 times that employed in permeability tests.77, 78

51Cr-EDTA has also been criticized for being a radioactive pharmaceutical. The estimated radiation dose of a 100 μCi (3.7 Mbq) dose is 0.12 mSv. In comparison, a chest X-ray gives 0.05 mSv, abdominal X-ray 1.4 mSv and the total radiation dose from natural sources is about 2 mSv/year.43 The effective dose, therefore, is equivalent and comparable to other standards currently routinely employed in nuclear medicine.


  1. Top of page
  2. Abstract
  10. References


More recently, sucrose has been introduced as a marker of NSAID-induced gastroduodenal damage.79, 80 Sucrose permeability, unlike endoscopy, is cheap, simple and readily accepted by patients. Increased urinary excretion of sucrose after an oral dose indicates abnormalities in the epithelium of the gastroduodenum. When entering the lower intestine, sucrose is readily cleaved to monosugars due to sucrase activity in the brush border membrane, which makes sucrose unsuitable for assessing intestinal permeability because absorption of intact sucrose is restricted to the upper gastroduodenal mucosa. Hence, detection of sucrose in the urine indicates leakage of the GI segments proximal to sucrase enzyme activity (i.e. stomach and duodenum). A study using balloon pyloric occlusion has independently confirmed that the major site of increased sucrose permeability is indeed the stomach.81

A test solution containing 100 g of sucrose in 350 mL of water is consumed at bedtime after at least a 3-h fast. All urine is collected from the subjects in a pre-weighed container containing 5 mL of 10% thymol in methanol as a preservative.79

It has been demonstrated that sucrose permeability correlates with severity of the upper gastroduodenal damage and increases with repetitive exposure of NSAIDs.79[80][81][82][83]–84 Increased sucrose permeability in man may be useful in predicting the presence of clinically significant gastric disease seen upon endoscopy.80 The sensitivity for detecting mild gastritis, duodenitis, severe gastritis and gastric ulcers was 16, 29, 69 and 84%, respectively.80 The specificity for predicting an abnormal endoscopy was 96%.80 Another study has reported that for mild gastritis, duodenitis and severe gastritis the sensitivity for detection was 17, 29 and 68%, respectively; no gastric ulcers were seen in this latter study.81

Animal studies in dogs indicate healing of gastric epithelial damage can also be monitored through the use of sequential measurements of sucrose permeability.83 Furthermore, the sucrose probe has been shown to be able to detect differences in both NSAID formulation and dose.85, 86 Moreover, economic analysis of the sucrose test indicates that sucrose permeability is an attractive means of stratifying elderly RA patients who are at risk of developing complications from chronic NSAID use.87

Patients with oesophagitis did not have elevated sucrose permeability compared to controls; this may be due to the small oesophageal surface area and a short contact time with the diseased mucosa.80 In addition, duodenal disease was erratically predicted by increased sucrose permeability, which may be due to rapid sucrose degradation within the duodenum, the short contact time of sucrose solution within the affected duodenal tissue and sucrase activity.80 Further studies that modify the absorption rate of the sucrose test solution could possibly be adaptable to detection of severe oesophagitis or duodenal disease.

As indicated in the preliminary manuscript by Meddings et al.,79 sucrose is a suitable and selective probe for the gastroduodenum in a healthy intestine without disaccharide deficiency. Any disease state or xenobiotic which affects sucrase activity, may affect the interpretation of the permeability data and is a limitation to its diagnostic use. Indeed, studies have utilized alterations in sucrose urinary excretion for non-invasive investigation of intestinal disaccharidase activity caused by α-glucosidase inhibition, primary hypolactasia and coeliac disease.88


  1. Top of page
  2. Abstract
  10. References

There are few published clinical studies on NSAID-induced gastroduodenal permeability owing to the novelty of the sucrose probe. However, this technique has been independently verified by several groups and has uniformly demonstrated the ability to detect permeability changes induced by acetylsalicylic acid and other NSAIDs.80[81][82]–83, 87, 89 Furthermore, randomized, prospective trials evaluating sucrose permeability in long-term NSAID users are still required to determine the potential clinical utility.

The frequency distribution of baseline permeability values in 65 healthy male volunteer humans from 0 to 24 h are also positively skewed with a median of 2.45% excretion (95% CI: 2.11–2.86).38 In humans, there is a good linear relationship between the 0–6 and 6–24 h cumulative excretion of 51Cr-EDTA.33 Furthermore, there does not appear to be any gender- or age-related differences in baseline permeability values in human studies.91, 92

The original study of NSAID-induced intestinal permeability was undertaken in healthy subjects after ingestion of two doses of aspirin (1.2 and 1.2 g), ibuprofen (400 and 400 mg) and indomethacin (75 and 50 mg) at midnight and 1 h before a 51Cr-EDTA permeability test the following day.17 Intestinal permeability increased significantly from control levels following each drug and, in this limited example, the effect was correlated with NSAID potency to inhibit cyclooxygenase in vitro. However, no estimation of systemic or tissue cyclooxygenase-1 or -2 inhibition was attempted in this study. Intestinal permeability also increased to a similar extent after oral and rectal administration of indomethacin, suggesting that this effect of NSAIDs may be systemically mediated and/or that biliary excretion is of importance.33

NSAIDs inhibit cyclooxygenase and therefore inadequate prostaglandin generation is derived from fatty acids by the damaged cell membranes, contributing to GI ulceration.15 Bjarnason and co-workers first examined the influence of concomitant prostaglandin administration on indomethacin-induced permeability changes with prostaglandin E2, a naturally occurring prostaglandin, which did not seem to reverse permeability changes. This lack of effect was postulated to be due to the instability of the preparation. However, prostaglandin E2 itself significantly decreased baseline permeation of 51Cr-EDTA.17 Misoprostol (a prostaglandin E1 analogue) alone demonstrated little or no effect on permeation of 51Cr-EDTA,63, 90 however, coadministered with indomethacin, misoprostol in high doses for short periods protected the small bowel mucosa from the effects of indomethacin, as reflected by decreasing indomethacin-induced increased intestinal permeation of 51Cr-EDTA. Rioprostol, a prostaglandin analogue given in small doses at the same time as indomethacin, has a maximally protective action, however, the excretion values were still significantly higher than baseline.43 The permeation of other markers (i.e. 3-O-methyl glucose, D-xylose and L-rhamnose) was not affected by indomethacin. Conversely, a lack of reduction of indomethacin-induced intestinal permeability with misoprostol has subsequently been demonstrated.90 In this study only 800 μg misoprostol was administered every 24 h for 150 mg doses of indomethacin. In addition, the NSAID and the prostaglandins were administered at different times in this study so that the intestinal protection may not have been evident at the site of absorption, whereas in the initial study 1200 μg misoprostol was administered over 20 h with 125 mg indomethacin given during the last 12 h.17 These studies therefore suggest that the protective effects of misoprostol against indomethacin-induced intestinal permeability may be dose-dependent and/or that intestinal permeability may only be partially mediated by reduced mucosal prostaglandins.

Sulphasalazine and metronidazole have been clinically used to reduce NSAID-induced enteropathy.93, 94 The effects of indomethacin treatment (50 mg three times a day) for 1 week with coadministered metronidazole 400 mg twice a day significantly reduced indomethacin-induced permeability changes.90 This may suggest a role for intestinal bacteria in the pathogenesis of the initial permeability increase induced by NSAIDs. A two-stage process in NSAID-induced intestinal damage has been suggested.90 Initially, there is superficial and reversible mucosal damage, probably directly related to NSAID effects on local prostaglandin depletion but independent of luminal contents. The second stage is independent of direct drug effects but follows if initial damage to mucosal defences is not prevented or reversed. Due to the increased epithelial permeability, harmful luminal contents, including bacterial flora, have increased access to the GI mucosa and this may pave the way for more severe GI damage and the development of NSAID enteropathy.

The ability of the 51Cr-EDTA test to detect an increase in intestinal permeability, which resulted from administration of two different doses of naproxen, has been reported.71 There was a statistically significant difference between the median increase as a percentage of the baseline excretion for 750 mg naproxen (19%), and 1000 mg naproxen (68%). Misoprostol did not appear to protect against naproxen-induced increased intestinal permeability.95 However, the possibility of a dose-dependent protective effect upon higher doses of misoprostol was not examined. In addition, the small sample numbers involved in this study (n = 6) may have been inadequate for firm conclusions from these results. The protective effects of sucralfate against NSAID-induced GI damage have also been shown to be confined to the upper gastroduodenum, because sucralfate did not provide protection from naproxen-induced permeability changes in the small intestine.96

Some NSAIDs are formulated as prodrugs that are inactive as COX inhibitors until after absorption; it has been suggested that they might cause less intestinal damage than other NSAIDs.97 After a 1-week treatment period of sulindac 200 mg daily there was no apparent increase in intestinal permeability above baseline, whereas an indomethacin treatment of 2 mg/kg/day in three divided doses increased 51Cr-EDTA permeation. Similar results have recently been reported with another prodrug NSAID, nabumetone.65 After taking nabumetone, 1 g at midnight for 7 days, nabumetone did not appear to induce an increase in 51Cr-EDTA permeability above baseline, whereas indomethacin treatment significantly enhanced 51Cr-EDTA permeation. These results could be interpreted to suggest that the systemically mediated effect of NSAIDs is weak and that the main damage is sustained after drug absorption or excretion in bile. However, a major contribution to local effects of NSAIDs in inducing permeability changes contradicts a previous assertion of a major systemic effect being responsible for NSAID-induced intestinal permeability after rectal administration of NSAIDs.33 These results may also be attributed to either intestinal permeability being reflective of mechanisms other than reduction of prostaglandins such as diversion of arachidonic acid metabolism down the lipoxygenase pathway, and oxy-radical production or direct drug cytotoxicity, or the differing pattern of biliary excretion of nabumetone (0%), sulindac (4%) and indomethacin (36%).98, 99 In addition, the sample numbers in these studies are small and no dose–response relationship with this effect were mentioned. Moreover, in the Bjarnason study65 the collection period was only 5 h. Aabakken suggests that a 24 h collection period increases the sensitivity of the test.38 Furthermore, there was a trend for nabumetone to increase permeability from baseline in several subjects, which may reach statistical significance with a larger sample size. A recent abstract from a separate laboratory has suggested that nabumetone increased intestinal permeability to the same degree as indomethacin.100 However, these studies all used different methods for assessing intestinal permeability, which vary in their sensitivity; this may account for the apparent discrepancies.

In addition to inhibiting prostaglandin synthesis it has been suggested that NSAIDs may inhibit glycolysis and the tricarboxylic acid cycle resulting in inhibition of oxidative phosphorylation and damage to the enterocyte resulting in a depletion of cellular ATP.17, 67 The consequence of reduced ATP production is a collapse of the cytoskeleton and, therefore, intercellular tight-junction regulation is disrupted. Indomethacin administered as 50 or 75 mg increased intestinal permeability, whereas a formulation of indomethacin containing 15 mg glucose and 15 mg citrate to each milligram of indomethacin prevented an increase in intestinal permeability above baseline values.67 Unfortunately, no further clinical studies have validated these claims. Conversely, the NSAID azapropazone, through its incorporation into a glucose/citrate formulation in proportions of 1:1:1, demonstrated no significant reduction in GI microbleeding.101

In a validated rat model of gastrointestinal permeability,72, 85, 102, 103 the ameliorative effect of orally administered glucose/citrate on indomethacin-induced intestinal permeability was examined after subcutaneous administration of indomethacin and oral administration of glucose/citrate.104 The lack of protection of glucose/citrate given orally when indomethacin is administered subcutaneously suggests a possible presystemic and, perhaps, physicochemical interaction. No reduction in the indomethacin bioavailability in humans was evident upon glucose/citrate administration.67 In this pharmacokinetic study the indomethacin and glucose/citrate formulation was ingested with 100 mL of water, whereas in their intestinal permeability study, the indomethacin and glucose/citrate was administered with only 50 mL of water, which might have increased the dissolution of the formulation in the pharmacokinetic study. In Bjarnason’s pharmacokinetic study there was a 5 and 17% reduction in the AUC and Cmax for the indomethacin and glucose/citrate formulation, respectively. This difference was, however, not statistically significant. In Bjarnason’s permeability study the 51Cr-EDTA excretion, although significantly higher for indomethacin alone, gave only small differences between the two formulations; 1.15 ± 0.15% vs. 1.54 ± 0.19% for indomethacin and glucose/citrate (1:15:15) control and indomethacin treated subjects, respectively. Bjarnason et al.15 have more recently indicated that this cytoprotective effect of indomethacin and glucose/citrate to prevent permeability changes and enteropathy is not evident upon repetitive administration of the indomethacin and glucose/citrate formulation.17 It is not stated, however, if this new study found any cytoprotective effect after a single indomethacin and glucose/citrate dose. If there is no long-term protection then it is likely that biliary secretion and systemic concentrations are responsible for this effect. Further independent studies are required to reconcile these discrepancies.

Although the deleterious effects of NSAIDs in the intestine are becoming more common, very few studies have examined the possible protective measures in the distal intestine. Furthermore, differentiation between the gastric and intestinal manifestations of NSAIDs has been largely ignored. Hence, it is not known if the suggested schemes for upper gastroduodenal protection also alleviates NSAID-induced intestinal lesions. Aabakken’s group have examined the use of the H2-antagonists (cimetidine and famotidine) and sucralfate, none of which seem to possess any apparent protective or antagonizing effect on naproxen-induced intestinal permeability.105, 106

In addition, in a latin-square crossover study, naproxen 500 mg twice daily for 7 days as plain tablets, enteric-coated tablets or enteric-coated granules in capsules was administered to healthy male volunteers. All formulations induced a significant increase in 51Cr-EDTA permeability, but no statistical differences were detected between them. A considerable inter-individual variation was seen; however, it appears that the median urinary excretion values for the enterogranulate capsules and the enterocoated tablets were higher than after plain tablets.107 The effect of modified-release formulations on NSAID-induced permeability has not been adequately addressed in other clinical studies. The results of the intestinal permeability of the sustained-release product appear to be more variable than the regular release product, possibly due to the sustained release of drug and its continuous presence within the intestinal tract which may induce more local distal intestinal damage in addition to the systemically mediated effects. These data are consistent with those reported by Choi et al.,108 who found increased intestinal permeability with a sustained- release formulation of diclofenac but did not see a statistically significant increase in intestinal permeability with regular release diclofenac. These observations also appear to agree with clinical observations of distal intestinal damage induced by sustained release and enterocoated NSAIDs. Evidence for pre-systemic distal intestinal damage has come from the osmotically activated slow-release formulation of indomethacin (Osmosin; Merck-Sharpe and Dohme, UK) no longer commercially available. Osmosin capsules were located at the site of perforating colonic and ileal ulcers and free in the peritoneal cavity.109 Furthermore, a more recent report has suggested the location of possible diclofenac pill fragments at the site of ulceration and strictures.110 This may lead to high local concentrations in the ileum and colon, inducing distal GI damage by NSAIDs.

Although the more distal intestinal manifestations of sustained-release NSAID formulations have been largely ignored, the likelihood of its increased occurrence with more frequent use of NSAID medication has been previously predicted.111 Cost containment of pharmaceuticals is topical and the therapeutic rationale behind enteric-coated and sustained-release formulations in terms of GI side-effects is not clear-cut. These data suggest that sustained-release NSAIDs do not solve the problem of NSAID-induced GI toxicity but merely shift the problem to a more distal site within the GI tract.

Mielants et al.91 report that there was no significant difference in intestinal permeability between patients taking NSAIDs and patients taking corticosteroids. This further suggests that alteration of intestinal permeability may not only be accounted for by an inhibition of mucosal COX activity but that other pathways in the arachidonic acid cascade might be implicated.

It also appears that the effect of NSAIDs on increasing intestinal permeability differs markedly from their potency to cause gastric damage. Aspirin has been shown to induce gastroduodenal permeability changes79, 82, 83 and is well-known to induce upper gastroduodenal damage. However, intestinal damage for aspirin, measured as increased intestinal permeability, has been shown to be minimal.17, 88 Conversely, it has also been suggested that aspirin increased intestinal permeability in a limited number of patients. However, the type of formulation and dose administered was not mentioned in this latter study.112 The lack of effect of aspirin on intestinal permeability may be due to the rapid absorption of aspirin from the upper part of the gastroduodenum and its efficient hydrolysis to salicylic acid and lack of enterohepatic recirculation which limits both direct exposure of the more distal intestine to the drug and the availability of aspirin for systemic distribution into the intestinal mucosa. Salicylic acid is a very weak inhibitor of COX and has also been shown to be ineffective in inducing leucocyte adherence to the vascular endothelium.113 The concentrations required for 50% inhibition of COX1 have been found to be 0.3 and 35 mg/L for aspirin and salicylic acid, respectively.114

Omega-3 fatty acid (fish oil) ingestion inhibits both neutrophil LTB4 release and production of platelet activating factor, which have been implicated in the pathogenesis of gastrointestinal ulceration.115 The potential protective effects of omega-3 fatty acid ingestion in subjects consuming NSAIDs did not demonstrate any significant differences in endoscopic gradings of visible changes in either the stomach or duodenum and small intestinal permeability changes as assessed by 51Cr-EDTA also showed no difference between patients consuming fish oil or corn oil.116

It has been postulated that there may be racial differences in susceptibility to NSAID adverse reactions. In 12 healthy volunteers of Afro-Caribbean (AC), Asian (As) and Caucasian (Cs) origin the baseline differential excretion of lactulose/ L-rhamnose was 0.035 ± 0.004, 0.12 ± 0.046, 0.058 ± 0.013 in the Cs, As and AC groups, respectively, with a significant difference between the AC and Cs groups. Following two 75 mg doses of indomethacin, the increase was significant in the Cs and AC groups. This initial study suggests that there is a difference in intestinal permeability between racial groups. The non-Caucasians had increased permeability which may be the basis of racial variation in adverse effects to NSAIDs.117 Further studies are obviously required in the light of these findings.

Unfortunately, due to the ethical constraints of repeated administration of radiolabelled compounds in humans and the technical difficulty of access to the human distal intestine these studies in humans have been performed in the absence of any pharmacokinetic considerations and, thus, neither the time-course of these changes nor their relationship to drug concentration in plasma or GI tissues have been established. Furthermore, complete characterization of the dose–effect relationship for these NSAIDs has not been described due to maximum ethical daily dosage regulations. However, it appears that the relative effect of indomethacin on intestinal permeability is short-lived, with restoration of intestinal integrity within a week of the last ingested dose.65 This contrasts with that seen in NSAID enteropathy where inflammation may persist for over 16 months after stopping NSAIDs.21


  1. Top of page
  2. Abstract
  10. References

51Cr-EDTA, unlike lactulose, is resistant to degradation by bacteria in the lower intestine.61, 73 The permeation of ingested 51Cr-EDTA is higher than lactulose in normal adults, however, there are no significant differences in patients with ileostomies.61, 73 It has been suggested that the simultaneous administration of lactulose and 51Cr-EDTA may enable changes in colonic permeability to be distinguished from changes in the small intestine.73, 118

Further advances in gastrointestinal permeability detection are still possible in both the basic and clinical sciences. Recently, Meddings et al.119, 120 have demonstrated the basic and clinical utility of sucralose as a non-degradable sugar and as a suitable non-invasive marker of colonic disease which appears to correlate with disease severity. Utilizing sucrose, lactulose/mannitol and sucralose probes, respectively, non-invasive detection of gastric, enteric or colonic damage in a single test can be made. Sucralose is an attractive alternative to 51Cr-EDTA for measuring colonic disease, because the latter radioactive probe possesses ethical restrictions for routine or serial tests, especially in children. Either oral or rectal administration of sucralose could be used and the formulation of a sucralose enema or foam is an intriguing concept that could conceivably be instilled in patients during a colonoscopy procedure to site-specifically evaluate colonic permeability.

Although rare, NSAIDs can also induce mucosal damage in the large intestine.16 In addition, NSAIDs and selective COX2 inhibitors can cause reactivation of inflammatory bowel disease.16, 121 The development of selective probes of colonic permeability may have both basic and clinical applications in the study of NSAID toxicity in the large bowel. No clinical studies to date have assessed colonic permeability in patients with inflammatory bowel disease who are taking NSAIDs.


  1. Top of page
  2. Abstract
  10. References

Gastrointestinal permeability tests have demonstrated utility in both basic and clinical research in the investigation of various intestinal diseases including NSAID-induced alterations. An implicit advantage of permeability tests is that they can site-specifically reflect functional integrity over a major area of the intestinal mucosa and sequential permeability tests may allow non-invasive assessment of cytoprotection and prophylaxis approaches, as well as healing of the intestinal barrier. These tests are safe, well tolerated, reproducible and easy to perform and because of their non-invasive nature can easily be applied to diagnostic screening, research, and complement the use of invasive investigations of gastrointestinal disease such as radiology, biopsy and endoscopic procedures. Future research should determine whether the use of gastrointestinal permeability tests could prove cost-effective by reducing endoscopic and radiologic workload through screening and by selecting patients likely to benefit from further invasive procedures and to assess clinical progress and response to treatment. The acceptance of these techniques remains to be determined after their clinical reliability is evaluated by primary care physicians.


  1. Top of page
  2. Abstract
  10. References
  • 1
    Koutsos MI, Shiff SJ, Rigas B. Can nonsteroidal anti-inflammatory drugs be recommended to prevent colon cancer in high risk patients? Drugs Aging 1995; 6(6): 421 5.
  • 2
    Rogers J, Kirby LC, Hempelman SR, et al. Clinical trial of indomethacin in Alzheimer’s disease. Neurology 1993; 43: 1609 11.
  • 3
    Blower PR. Nonsteroidal anti-inflammatory drugs. Br J Rheumatol 1993; 32(Suppl. 4): 35 8.
  • 4
    Baum C, Kennedy DL, Forbes MB. Utilization of nonsteroidal antiinflammatory drugs. Arth Rheum 1985; 28: 686 92.
  • 5
    Epstein AM, Read JL, Winickoff R. Physician beliefs, attitudes, and prescribing behavior for anti-inflammatory drugs. Am J Med 1984; 77: 313 8.
  • 6
    Wilkens RF. The selection of a nonsteroidal antiinflammatory drug. Is there a difference? J Rheumatol 1992; 19(Suppl.36): 9 12.
  • 7
    Myers ABR. Salicin in acute rheumatism. Lancet 1876; 2: 676 7.
  • 8
    Rainsford KD. Introduction and historical aspects of the side-effects of anti-inflammatory analgesic drugs. In: Rainsford KD, Velo JP, eds. Side-effects of Anti-inflammatory Drugs. Part 1: Clinical and Epidemiological Aspects. Lancaster: MTP Press, 1987: 3–26.
  • 9
    Fries JF. NSAID gastropathy: Epidemiology. J Musculoskeletal Med 1991; 8(2): 21 8.
  • Committee on the Safety of Medicines.
    Non-steroidal anti-inflammatory drugs and serious gastrointestinal adverse reactions-2. Br Med J 1986; 292: 1190 1.
  • 11
    Giercksky K-E, Huseby G, Rugstad H-E. Epidemiology of NSAID-related gastrointestinal side effects. Scand J Gastroenterol 1989; 24(Suppl. 163): 3 8.
  • 12
    Silvoso G, Ivey KJ, Butt J. Incidence of gastric lesions in patients with rheumatic disease on chronic aspirin therapy. Ann Intern Med 1979; 91: 517 20.
  • 13
    Morris AJ, Wasson LA, Mackenzie JF. Small bowel enteroscopy in undiagnosed gastrointestinal blood loss. Gut 1992; 33: 887 9.
  • 14
    Aabakken L. Non-steroidal anti-inflammatory drugs—the extending scope of gastrointestinal side-effects. Aliment Pharmacol Ther 1992; 6: 143 62.
  • 15
    Bjarnason I, Hayllar J, Macpherson AJ, et al. Side effects of nonsteroidal anti-inflammatory drugs on the small and large intestine. Gastroenterology 1993; 104: 1832 47.
  • 16
    Davies NM. Toxicity of nonsteroidal anti-inflammatory drugs in the large intestine. Dis Colon Rectum 1995; 38(12): 1311 21.
  • 17
    Bjarnason I, Williams P, Smethurst P, et al. Effect of non-steroidal anti-inflammatory drugs and prostaglandins on the permeability of the human small intestine. Gut 1986; 27: 1292 7.
  • 18
    Langman MJS, Morgan L, Worall A. Use of antiinflammatory drugs by patients admitted with small or large bowel perforations and haemorrhage. Br Med J 1985; 290: 347 9.
  • 19
    Allison MC, Howatson AG, Torrance CJ, et al. Gastrointestinal damage associated with the use of nonsteroidal antiinflammatory drugs. N Engl J Med 1992; 327(11), 751–6.
  • 20
    Davies NM, Jamali F, Skeith KJ. Non-steroidal antiinflammatory drug (NSAID)-induced enteropathy and severe chronic anemia in a patient with rheumatoid arthritis. Arth Rheum 1996; 39(2): 321 4.
  • 21
    Bjarnason I, Zanelli G, Smith T, et al. Nonsteroidal antiinflammatory drug-induced intestinal inflammation in humans. Gastroenterology 1987; 93: 480 9.
  • 22
    Bjarnason I, Prouse P, Smith T, et al. Blood and protein loss via small intestinal inflammation induced by non-steroidal anti-inflammatory drugs. Lancet 1987; 2: 711 14.
  • 23
    Bjarnason I, Zanelli G, Smith T, et al. The pathogenesis and consequence of non-steroidal anti-inflammatory drug induced small intestinal inflammation. Scand J Rheumatol 1987; Suppl. 64: 55 62.
  • 24
    Bjarnason I, So A, Levi AJ, et al. Intestinal permeability and inflammation in rheumatoid arthritis: Effects of non-steroidal anti-inflammatory drugs. Lancet 1984; 2: 1171 3.
  • 25
    Rooney PJ, Jenkins RT, Smith KM, et al. 111Indium-labelled polymorphonuclear leucocytes scans in rheumatoid arthritis—an important clinical cause of false positive results . Br J Rheumatol 1986; 25: 167 70.
  • 26
    Mitchell JA, Akarasereenont P, Thiemermann C, et al. Selectivity of nonsteroidal antiinflammatory drugs as inhibitors of constitutive and inducible cyclooxygenase. Proc Natl Acad Sci USA 1993; 90: 11693 7.
  • 27
    Wallace JL, Reuter B, Cicala C, et al. Novel nonsteroidal anti-inflammatory drug derivatives with markedly reduced ulcerogenic properties in the rat. Gastroenterology 1994; 107: 173 9.
  • 28
    Lichtenberger LM, Wang ZM, Romero JJ, et al. Non-steroidal anti-inflammatory drugs (NSAIDs) associate with zwitterionic phospholipids: insights into the mechanism and reversal of NSAID-induced gastrointestinal injury. Nat Med 1995; 1: 154 8.
  • 29
    Jaszewski R. Frequency of gastroduodenal lesions in asymptomatic patients on chronic aspirin or non-steroidal anti-inflammatory drug therapy. J Clin Gastroenterol 1990; 12(1): 10–13.
  • 30
    Pounder R. Silent peptic ulceration: Deadly silence or golden silence. Gastroenterology 1989; 96: 626 31.
  • 31
    Walt RP. Misoprostol for the treatment of peptic ulcer and antiinflammatory-drug-induced gastroduodenal ulceration. N Engl J Med 1992; 327(22): 1575 80.
  • 32
    McCarthy DM. Nonsteroidal antiinflammatory drug-induced ulcers: management by traditional therapies. Gastroenterology 1989; 96: 647 55.
  • 33
    Graham DY, Argawal NM, Roth SH. Prevention of NSAID-induced gastric ulcer with misoprostol: multicentre, double-blind, placebo-controlled trial. Lancet 1988; 2: 1277 80.
  • 34
    Graham DY. Famotidine to prevent peptic ulcer caused by NSAIDs. New Engl J Med 1996; 335: 1322 [Letter].
  • 35
    Bjarnason I, Macpherson A, Hollander D. Intestinal permeability: an overview. Gastroenterology 1995; 108: 1566 81.
  • 36
    Travis S & Menzies I. Intestinal permeability: functional assessment and significance. Clin Sci 1992; 82: 471 88.
  • 37
    Tagesson C & Sjödahl R. Passage of molecules through the wall of the gastrointestinal tract. Urinary recovery of different-sized polyethylene glycols after intravenous and intestinal deposition. Scand J Gastroenterol 1984; 19: 315 20.
  • 38
    Aabakken L. 51Cr-Ethylenediaminetetraacetic acid absorption test: methodological aspects . Scand J Gastroenterol 1989; 24: 351 8.
  • 39
    Juby LD, Rithwell J, Axon ATR. Lactulose/mannitol test: an ideal screen for celiac disease. Gastroenterology 1989; 96: 79 85.
  • 40
    Csaky TZ. Factors involved in the integrated mechanism of intestinal absorption. In: Advances in Physiological Sciences. Proceedings of the 28th International Congress of Physiological Sciences, Budapest. Oxford: Pergamon, 1981; 22–33.
  • 41
    Chadwick VS, Phillips SF, Hoffman AF. Measurements of intestinal permeability using low molecular weight polyethylene glycols (PEG 400) I. Chemical analysis and biological properties of PEG 400. Gastroenterology 1977; 73: 241 6.
  • 42
    Cooper BT. Small intestinal permeability in clinical practice. J Clin Gastroenterol 1984; 6: 499 501.
  • 43
    Bjarnason I, Peters TJ, Levi AJ. Intestinal permeability: Clinical correlates. Dig Dis 1986; 4: 83 92.
  • 44
    Smith MD & Brooks PM. Increased bowel permeability in ankylosing spondylitis and active RA. Aust NZ J Med 1984; 14(Suppl. 1): 345 6.
  • 45
    Parke AL, Fagan EA, Chadwick VS, et al. Gut permeability in Sjögren’s syndrome, RA, and coeliac disease. Ann Rheum Dis 1983; 42: 216.
  • 46
    Maxton DG, Bjarnason I, Reynolds AP, et al. Lactulose 51Cr-labelled ethylenediaminetetra-acetate, L-rhamnose and polyethylene glycol 500 as probe markers for assessment in vivo of human intestinal permeability . Clin Sci 1986; 71: 71 80.
  • 47
    Fälth-Magnuson K, Jansson G, Stenhammar L, et al. Intestinal permeability assessed with different-sized polyethylene glycols in children undergoing small-intestinal biopsy for suspected celiac disease. Scand J Gastroenterol 1989; 24: 40 6.
  • 48
    Peters TJ & Bjarnason I. Uses and abuses of intestinal permeability measurements. Can J Gastroenterol 1988; 2(3): 127 32.
  • 49
    Ukabam SO & Cooper BT. Small intestinal permeability to mannitol, lactulose and polyethylene glycol 400 in coeliac disease. Dig Dis Sci 1984; 29: 809 16.
  • 50
    Skoldstam L, Larsson L, Lindstrom FD. Effects of fasting and lactovegetarian diet on rheumatoid arthritis. Scand J Rheumatol 1979; 8: 249 55.
  • 51
    Sundquist T, Magnusson KE, Sjödahl R, et al. Passage of molecules through the wall of the gastrointestinal tract. II. Application of low-molecular weight polyethylene glycol and a deterministic mathematical model for determining intestinal permeability in man. Gut 1980; 21: 208 14.
  • 52
    Hollander D, Vadheim CM, Bretholz E, et al. Increased intestinal permeability in patients with Crohn’s disease and their relatives. Ann Intern Med 1986; 105: 883 5.
  • 53
    Magnusson K-E, Sundquist T, Sjödahl R, et al. Altered intestinal permeability to low-weight polyethylene glycols (PEG 400) in patients with Crohn’s disease. Acta Chir Scand 1983; 149: 323 7.
  • 54
    Kendall MJ & Hawkins CF. Xylose test; effect of aspirin and indomethacin. Br Med J 1971; 1(5748): 533–6.
  • 55
    Dyer NH, Kendall MJ, Hawkins CF. Malabsorption in rheumatoid arthritis. Ann Rheum Dis 1971; 30: 626 30.
  • 56
    Struthers GR, Andrews DJ, Gilson RJR, et al. Intestinal permeability. Lancet; 1985; 2: 587 8.
  • 57
    Bjarnason I, Peters TJ, Veall N. 51Cr-EDTA test for intestinal permeability . Lancet 1984; 2: 523.
  • 58
    Hamilton I, Cobden I, Rothwell J, et al. Intestinal permeability in coeliac disease. The response to gluten withdrawal and single dose gluten challenge. Gut 1982; 23: 202 10.
  • 59
    Behrens RH, Szaz KF, Northrop C, et al. Radionuclide tests for assessment of intestinal permeability. Eur J Clin Invest 1987; 17: 100 5.
  • 60
    Bjarnason I, Peters TJ, Veall N. A persistent defect in intestinal permeability in coeliac disease demonstrated by a 51Cr-labeled EDTA absorption test . Lancet 1983; 1: 323 5.
  • 61
    Jenkins AP, Trew DR, Crump BJ, et al. Do non-steroidal anti-inflammatory drugs increase colonic permeability? Gut 1991; 32: 66 9.
  • 62
    O’mahoney S & Ferguson A. Small intestinal mucosal protection mechanisms and their importance in rheumatology. Ann Rheum Dis 1991; 50: 331 6.
  • 63
    Bjarnason I, Smethurst P, Fenn CG, et al. Misoprostol reduces indomethacin-induced changes in human small intestinal permeability. Dig Dis Sci 1989; 34: 407 11.
  • 64
    Bjarnason I. Experimental evidence of the benefit of misoprostol beyond the stomach in humans. J Rheumatol 1990; 17(Suppl. 20): 38 41.
  • 65
    Bjarnason I, Fehilly B, Smethurst P, et al. Importance of local versus systemic effects of non-steroidal anti-inflammatory drugs in increasing small intestinal permeability in man. Gut 1991; 32: 275 7.
  • 66
    Bjarnason I, Smethurst P, Levi AJ, et al. The effect of polyacrylic acid polymers on small-intestinal function and permeability changes caused by indomethacin. Scand J Gastroenterol 1991; 26: 685 8.
  • 67
    Bjarnason I, Smethurst P, Macpherson A, et al. Glucose and citrate reduce the permeability changes caused by indomethacin in humans. Gastroenterology 1992; 102: 1546 50.
  • 68
    Bjarnason I, O’morain C, Levi AJ, et al. Absorption of 51chromium-labeled ethylenediaminetetraacetate in inflammatory bowel disease . Gastroenterology 1983; 85: 318 22.
  • 69
    Lökken P & Sögnen E. 51Cr-EDTA as a reference substance in research on gastrointestinal functions . Acta Pharmacol Toxicol 1967; 25(4): 39.
  • 70
    Simpson LO. NSAID and the leaky gut. Lancet 1985; 1: 218 19.
  • 71
    Aabakken L & Osnes M. 51Cr-Ethylenediaminetetraacetic acid absorption test. Effects of naproxen, a non-steroidal, antiinflammatory drug . Scand J Gastroenterol 1990; 25: 917 24.
  • 72
    Reuter BK, Davies NM, Wallace JL. NSAID-enteropathy in rats: Role of epithelial permeability, bacterial load, and enterohepatic circulation. Gastroenterology 1997; 112(1): 109 17.
  • 73
    Elia M, Behrens R, Northrop C, et al. Evaluation of mannitol, lactulose and 51Cr-labeled ethylenediaminetetra-acetate as markers of intestinal permeability in man . Clin Sci 1987; 73: 197 204.
  • 74
    Mielants H & Vey EM. NSAID and the leaky gut. Lancet 1985; 1: 218.
  • 75
    Peled Y, Watz C, Gilat T. Measurement of intestinal permeability using 51Cr-EDTA . Am J Gastroenterol 1985; 80(10): 770 3.
  • 76
    Bjarnason I, Ward K, Peters TJ. The leaky gut of alcoholism: possible route of entry for toxic compounds. Lancet 1984; 1: 179 82.
  • 77
    Aronson AL & Rogers KM. Effects of calcium and chromium chelates of ethylenediaminetetraacetate on intestinal permeability and collagen metabolism in the rat. Toxicol Appl Pharmacol 1972; 21: 440 53.
  • 78
    Tibdall CS. Magnesium and calcium as regulators of intestinal permeability. Am J Physiol 1964; 206: 243 6.
  • 79
    Meddings JB, Sutherland LR, Byles NI, et al. Sucrose: A novel permeability marker for gastroduodenal disease. Gastroenterology 1993; 104: 1619 26.
  • 80
    Sutherland LR, Verhoef M, Wallace JL, et al. A simple, non-invasive marker of gastric damage: Sucrose permeability. Lancet 1994; 343: 998 1000.
  • 81
    Rabassa AA, Goddame R, Sutton FM, et al. Effect of aspirin and Helicobacter Pylori on the gastroduodenal mucosal permeability to sucrose. Gut 1996; 39: 159 63.
  • 82
    Irvine EJ, Goodacre RL, McCullough HN, et al. Damaged surface area of gastric mucosa correlates with urinary sucrose excretion. Arth Rheum 1995; 38(9): A57, S162.
  • 83
    Hildsen RJ, Meddings JB, Sutherland LR. Intestinal permeability changes in response to acetylsalicylic acid in relatives of patients with Crohn’s disease. Gastroenterology 1986; 110: 1395 403.
  • 84
    Meddings JB, Kirk D, Olson ME. Non-invasive detection of nonsteroidal anti-inflammatory drug-induced gastropathy in dogs. Am J Vet Res 1995; 56(8): 977 81.
  • 85
    Davies NM, Corrigan BW, Jamali F. Sucrose urinary excretion in the rat measured using a simple assay: a model of gastroduodenal permeability. Pharm Res 1995; 11: 1733 6.
  • 86
    Davies NM & Jamali F. Influence of dosage form on the gastroenteropathy of flurbiprofen in the rat: evidence of shift in the toxicity site. Pharm Res 1997 ; 14(11): 1597 1600.
  • 87
    Steinhart AH, Detsky AS, Bombardier C. Economic analysis of the sucrose test: a non-invasive test of gastroduodenal mucosal damage. Gastroenterology 1996; 110(4): A41(Abstract).
  • 88
    Bjarnason I, Batt R, Catt S, et al. Evaluation of differential disaccharide excretion in urine for non-invasive investigation of altered intestinal disaccharidase activity caused by α-glucosidase inhibition, primary hypolactasia and coeliac disease. Gut 1996; 39: 374 81.
  • 89
    Ryan AJ, Chang R-RT, Gisolfi CV. Gastrointestinal permeability following aspirin intake and prolonged running. Med Sci Sports Exerc 1996; 28(6): 698 705.
  • 90
    Davies GR, Wilkie ME, Rampton DS. Effects of metronidazole and misoprostol on indomethacin-induced changes in intestinal permeability. Dig Dis Sci 1993; 38(3): 417 25.
  • 91
    Mielants H, Goemaere S, De Vos M, Schelstraete K, et al. Intestinal mucosal permeability in inflammatory rheumatic diseases. I. Role of antiinflammatory drugs. J Rheumatol 1991; 18: 389 93.
  • 92
    Mielants H, De Vos M, Goemaere S, et al. Intestinal mucosal permeability in inflammatory rheumatic diseases. II. Role of disease. J Rheumatol 1991; 18: 394 400.
  • 93
    Bjarnason I, Hayllar J, Smethurst P, et al. Metronidazole reduces intestinal inflammation and blood loss in non-steroidal anti-inflammatory drug induced enteropathy. Gut 1992; 33: 1204 8.
  • 94
    Hayllar J, Smith T, Macpherson A, et al. Nonsteroidal anti-inflammatory drug-induced small intestinal inflammation and blood loss. Arth Rheum 1994; 37(8): 1146 50.
  • 95
    Jenkins RT, Rooney PJ, Hunt RH. Increased bowel permeability to [51Cr]EDTA in controls caused by Naproxen is not prevented by cytoprotection . Arth Rheum 1988; 31(Suppl. 1): R11.
  • 96
    Aabakken L, Larsen S, Osnes M. Sucralfate for prevention of naproxen-induced mucosal lesions in the proximal and distal gastrointestinal tract. Scand J Rheumatol 1989; 18: 361 8.
  • 97
    Davies GR & Rampton DS. The pro-drug sulindac may reduce the risk of intestinal damage associated with the use of conventional non-steroidal anti-inflammatory drugs. Aliment Pharmacol Ther 1991; 5: 593 8.
  • 98
    Dujovone CA, Pitterman A, Vincek WC, Drorinska MR, Davies RO, Duggan DE. Enterohepatic circulation of sulindac and metabolites. Clin Pharmacol Ther 1983; 33: 172 7.
  • 99
    Helleberg L. Clinical pharmacokinetics of indomethacin. Clin Pharmacokin 1981; 6(4): 245 58.
  • 100
    Devlin J, Moots R, Andrews D, Boivin C, Emery P. High intestinal permeability in patients with inflammatory arthritis is more strongly associated with any NSAID therapy than with diagnosis or disease activity. Arth Rheum 1996; S280: 1513.
  • 101
    Rainsford KD, Dieppe PA, Pritchard MH, et al. Protection from gastrointestinal side-effects of azapropazone by its incorporation into a glucose sodium acid citrate formulation. Aliment Pharmacol Ther 1991; 5(4): 419 33.
  • 102
    Davies NM, Wright MR, Jamali F. Antiinflammatory drug-induced small intestinal permeability: rat is a suitable model. Pharm Res 1994; 11: 1564 9.
  • 103
    Ford J, Martin SW, Houston JB. Assessment of intestinal permeability changes induced by nonsteroidal anti-inflammatory drugs in the rat. J Pharm Toxicol Meth 1995; 34: 9 16.
  • 104
    Davies NM, Wright MR, Jamali F. Glucose/Citrate (G/C) gastrointestinal cytoprotective effect: a physiochemical G/C–NSAIDs interaction. Pharm Res 1995; 12(9): A8122, S357.
  • 105
    Aabakken L, Larsen S, Osnes M. Cimetidine tablets or suspension for the prevention of gastrointestinal mucosal lesions caused by non-steroidal anti-inflammatory drugs. Scand J Rheumatol 1989; 18: 369 75.
  • 106
    Aabakken L, Bjørnbeth BA, Weberg R, et al. NSAID-associated gastroduodenal damage—does famotidine protection extend into the mid and distal duodenum? Aliment Pharmacol Ther 1990; 4: 295 303.
  • 107
    Aabakken L, Bjørnbeth BA, Hofstad B, et al. Comparison of the gastrointestinal side effects of naproxen formulated as plain tablets, enteric-coated tablets, or enteric-coated granules in capsules. Scand J Gastroenterol 1989; 24(Suppl. 163): 65 73.
  • 108
    Choi VMI, Coates JE, Chooi J, et al. Small bowel permeability—a variable effect of NSAIDs. Clin Invest Med 1995; 18(5): 357 61.
  • 109
    Day TK. Intestinal perforation associated with osmotic slow release indomethacin capsules. Br Med J 1983; 287(6406): 1671 2.
  • 110
    Whitcomb DC. Pathophysiology of nonsteroidal antiinflammatory drug-induced intestinal strictures. Gastroenterology 1993; 105(5): 1590.
  • Anonymous.
    NSAIDs and gut damage. Lancet 1989; 2: 600.
  • 112
    Jenkins RT, Rooney RT, Jones DR, Bienenstock J, Goodacre RL. Increased intestinal permeability in patients with rheumatoid arthritis: a side-effect of oral non-steroidal antiinflammatory drug therapy? Br J Rheumatol 1987; 26: 103 7.
  • 113
    Asako H, Kubes P, Wallace JL, et al. Modulation of leukocyte adhesion in rat mesenteric venules by aspirin and salicylate. Gastroenterology 1992; 103: 146 52.
  • 114
    Mitchell JA, Akarasereenont P, Thiemermann C, Flower RJ, Vane JR. Selectivity of non-steroidal antiinflammatory drugs as inhibitors of constitutive and inducible cyclooxygenase. Proc Natl Acad Sci 1994; 90: 11693 7.
  • 115
    Sperling RI, Weinblatt M, Robin J-L. Effects of dietary supplemenation with marine fish oil on leukocyte lipid mediator generation and function in rheumatoid arthritis. Arth Rheum 1987; 30: 988 97.
  • 116
    Kremer JM, Malamood H, Maliakkal B, Rodgers JB, Ross JS, Cooper JA. Fish oil dietary supplementation for prevention of indomethacin induced gastric and small bowel toxicity in healthy volunteers. J Rheumatol 1996; 23: 1770 3.
  • 117
    Mahmud T, Scott DL, Crane R, Bjarnason I. NSAID-induced permeability. Change in whites and non-white ethnic subjects. Arth Rheum 1996; S280: 1514.
  • 118
    Jenkins AP, Nukajam WS, Menzies IS, et al. Simultaneous administration of lactulose and 51Cr-ethylenediaminetetraacetic acid . Scand J Gastroenterol 1992; 27: 769 73.
  • 119
    Ens R, Bak A, Meddings JB, Sutherland LR. Sucralose: A novel permeability marker for colonic disease. Can J Gastroenteol 1997; 11(Suppl. A): S37, 51A.
  • 120
    Meddings JB & Gibbons IR. Discrimination of site-specific alterations in gastrointestinal permeability in the rat. Gastroenterology 1998; 114(1): 83 92.
  • 121
    Reuter BK, Asfaha S, Buret A, Sharkey KA, Wallace JL. Exacerbation of inflammation-associated colonic injury in rat through inhibition of cyclooxygenase-2. J Clin Invest 1996; 98: 2076 85.