Continuing Education
Nonsteroidal anti-inflammatory drug gastropathy is it preventable?
-
P. M. Brooks Professor of Medicinc, Department of Medicine, I!niversiry of New South Wales, St Vincent’s IIospital, Sydney, NSW
N. D. Yeomans Professor of hledicine, Department of Medicine, University of Melbourne, Western Hospital, Melbourne, Vic
INTRODUCTION The non-steroidal anti-inflammatory drugs (NSAIDs) make up a group of the most commonly prescribed pharmaceuticals.‘ It is estimated that more than 30 million people throughout the world consume an NSAID each day.* T h e number of NSAID prescriptions written each year in the USA doubled between 1984 and 19883,4while in Australia the prevalence of NSAID use in elderly subjects has been estimated to be in excess of 20Y0.~T o a large extent this is a reflection of the high prevalence of rheumatic diseases in the community but an annual growth rate of 6% in NSAID prescriptions as seen in Canada during 1989 is unlikely to be due to increased disease incidence alone.’ As the population ages it is at risk from painful musculoskeletal diseases such as osteoarthritis, for which NSAIDs may not be the most appropriate treatment, and yet it seems that it is for these primarily non-inflammatory forms of arthritis that the majority of NSAIDs are prescribed.’ NSAIDs are associated with a wide spectrum of side effects including those involving renal, hepatic and haematological hnction and skin rashes. It is, however, the gastrointestinal (GI) side effects of NSAIDs which are most common and serious and make NSAIDs the most frequent cause of drug side effects reported to the Committee on Safety of Medicines in Britain each year.8 It is because of this that it is important to try to ascertain the true relationship between NSAID use and the clinically important gastrointestinal side effects and to develop strategies which might diminish this problem. In this review we will discuss briefly the pathophysiology of so-called ‘NSAID gastropathy’,
review the epidemiology of NSAID associated G I side effects and examine the clinical trial evidence that drugs might influence the ulcer healing rate and perhaps even prevent their development.
PATHOPHYSIOLOGY OF NSAID GASTROPATHY Mucosal damage in the gastrointestinal tract is the product of a fine balance between acid production and mucosal events. Surface epithelial cells secrete mucus and bicarbonateg and have the ability to rapidly repair injury.l0 These properties are essential for maintaining mucosal integrity in the stomach. Environmental factors such as smoking and stress, genetic predisposition, the bacterium Helicobacter pylori and NSAIDs (Figure 1) all play a role in predisposing to peptic ulceration. The epithelial cells can be damaged by acid, pepsin, refluxed bile acids, NSAIDs or the products of H. pylori. NSAIDs produce a wide spectrum of injury to the gastrointestinal tract ranging from petechial haemorrhages and erosions to ulcers. It is clear that the whole of the gastrointestinal tract is at risk from NSAID associated damage” with small and large bowel mucosal ulceration noted. NSAIDs probably have both a local and a systemic influence on G I mucosal physiology. Aspirin produces much of its damage topically, after it has entered the gastric surface mucus cells from the lumen. Evidence for this includes the high speed with which aspirin causes surface injury12 and the partial protection against injury which is provided by enteric coating of the aspirin.l3 However, the non-aspirin NSAIDs seem to produce most of their injury systemically, after they have been
Covresponiicwe: Professor 1’. M. Brooks, Ikpartment of Medicine, I’nlversiry of New South Wales, St Vincent’s Hospital, Darlinghurst, M W 2010, Australia. XONS’IEROIDAI, A N ’ I ‘ I - I ~ ~ ~ L A . M M A I OIIKCGS RY AND (;I
Aust NZ J Med 1992; 22
685
-
Arachidonic acid
MUCUS& HC03
NSAlDs
block synthesis
ndins
Microcirculation Figure I: Possible sites of action for the ‘cytoprotective’ acrions of endogenous prostaglandins in the gastric mucosa. I’rosragkdndins stimulate secrerion of mucus and bicarbonate by the rpithelial cells and also appear to stabilise the microcirculation.
absorbed. For example, in rat^,'^ parenteral administration of several commonly used NSAIDs produces at least as much gastric damage as does oral administration. Furthermore, studies in humans have shown that pro-drugs such as sulindac, enterically coated NSAIDs and rectally administered NSAIDs are all still associated with gastric damage - both acute and chr~nic,’’,’~albeit less frequently than with oral administration. It is possible that much of the gastric injury produced by NSAIDs is a consequence of the inhibition of the production of prostaglandins by the mucosal cells. This is caused by a blockade of cyclooxygenase at the start of the prostanoid biosynthetic pathway which is an effect common to all NSAIDs and, therefore, all NSAIDs share the capacity to produce gastric injury. There is now mounting evidence for several protective effects of gastric prostaglandins. They stimulate mucus and bicarbonate secretion from various cells” and seem to also increase the capacity of the mucosal surface to repel luminal acid - by increasing its hydrophobicity or ‘non-wettability’.18 Conversely, bicarbonate secretion is inhibited by several NSAIDS.’~ Moreover, severe gastric ulcers can be produced in experimental animals by inducing or administering antibodies to prostaglandins but not by antibodies to inactive prostaglandin analogues.20 Inhibition of prostaglandin production might enhance subsceptibility to damage by bile acids, hypertonic saline and even f o ~ d . * Prostaglandins ~-~~ might also stimulate 686 Aust NZ J Med 1992; 22
serosal-mucosal water flux so histodiluting injurious
agent^.'^ Some of the injurious effects of NSAIDs might be independent of blockade of prostaglandins. For instance, there is evidence that NSAIDs might be involved in the production of oxygen-free radicals.25 They may also impair angiogenesis at the base and margins of the ulcer, leading to ulcers which are more fibrous and less contractile than ~ s u a l . ~ ~ . ~ ’ T h e interaction between H. pylori and NSAIIh may also be important. However, evidence that these are independent risk factors for ulcers is mounting. One study in patients presenting with gastric ulcer found I€. pylori colonisation in nine of ten of the patients who were not taking NSAIDs but in only five of 12 of those taking the drugs.28 In a much larger study, Graham, et aL2’ found that the frequency of H. pylori infection was independent of NSAID use and that NSAID induced erosions were no more frequent in those with H. pylori infection than those without. It is, therefore, possible that there are two types of NSAID induced ulcers - ulcers that occur de ~ ~ o vino histologically normal gastric niucosa and ulcers that occur as a result of NSAID a d m i n i ~ t r a t i o n . ~ ~ T h e sequence of gastric mucosal events following ingestion of soluble aspirin has been reasonably well documented. Changes in the epithelial cells have been noted within minutes of aspirin ingestion,?‘ and by two hours nearly all subjects given a single dose will have red mucosal lesions and/or erosions. The subepithelial haemorrhages come and go during therapy with RKOOKS AND YEOl\l..lNS
NSAIDs.”*However, erosions occasionally fail to heal and progress to form ulcers. These, in turn, will often heal s p o n t a n e o ~ s l y It . ~ is ~ ~not ~ ~ clear, ~ however, in what proportion of patients this sequence of events is followed. For example, not all patients taking NSAIDs chronically will have erosions or ulcers when endoscoped. Silvoso, er a/., ’’ have demonstrated that patients taking eight aspirin tablets a day for at least three months show gastric erosions in 40% and gastric ulcers in 16%. This suggests that a large proportion of patients have adapted to NSAID challenge as suggested by Graham, er al. In these studies Graham showed that decreasing damage was apparent with continued exposure to aspirin and that lesions healed more rapidly if given a longer period of adaptive exposure to a ~ p i r i n . ”Adaptation ~ was more rapid at lower aspirin doses (1.3 gm per day) than higher doses (2.4 gm per day). This adaptation took several weeks to develop, during which time there was a gradual return to an almost normal mucosal appearance. Similar adaptation, although with a more rapid onset, has been seen in experimental animal^.^^,^' These experimental data are paralleled by epidemiological evidence. Carson, et a/.”x showed that the risk of gastrointestinal bleeding increased to a maximum after four NSAID prescriptions and decreased to no measurable risk after ten prescriptions. Griffin, et have shown that the maximal risk of gastrointestinal bleeding occurs in the first months after an NSAID prescription. O n the other hand, other groups have found no evidence of adaptation in humans using endoscopic assessment or measurement of gastrointestinal blood loss.“o,“l The majority of these studies have been done in patients who do not have chronic rheumatoid arthritis or osteoarthritis. There are, however, no clear data from the literature to suggest that there are differences between rheumatoid and non-rheumatoid gastric mucosa.
TABLE 1 _.__-___
9540 Ulct:r occurence
Ref
Reldtive risk
con‘idence interval
_ _ _ ~ _ _ _-~_ Gastric
Dmdenal
45 46
5 47
’3166
46
1 1
0 4-3 7
47 47
28 27
1554 1358
48 49 51
73 29 16
44’1 8 ’174 0737
39 39 53
42 79 29
1990 3 7-16 8 1 4 - 63
14769
Bleediyj
yastric tluotJerldl
Per‘oration peptic
(Jlcer dedtr
gastric duodenal pevtic
_ ~ _ _ _ _
-~
NSAID induced reaction per 10,000 person months of prescription^.^^ Using the technique of automated record linkage which relates diagnostic category to drug use incorporating information contained in large data bases, relative risks for perforated ulcer or haemorrhage as a result of NSAID therapy have been calculated to be around 1.S4Prospective case control studies, which probably give a more reliable estimate, indicate a substantially higher relative risk of serious upper G I side effects of between three and seven.j5 5 6 Fries et ul.,” using multivariate analysis in 2,400 patients with rheumatoid arthritis (RA) followed prospectively for an average period of 3.5 years, have demonstrated a hazard ratio (approximating to relative risk) for hospitalisations due to GI adverse effects of seven times that for patients not on NSAIDs. These data have led Fries et al. to suggest that in the United States the syndrome of NSAID induced EPIDEMIOLOGY gastropathy can be considered to account for at least The first point to remember is that diseases of the 2,600 deaths per annum and 20,000 hospitalisations gastric mucosa are extremely prevalent, with approxieach year in patients with RA alone. T h e studies of mately 10% of healthy asymptomatic people having Fries et al. suggest that those patients particularly at mucosal lesions at any one time and as many as 1% risk for development of NSAID related gastropathy showing extension into the submucosa, i.e., ulcer.42 are the elderly, those with a previous history of peptic Approximately 10% of us will experience frank peptic ulceration, those with severe rheumatoid disease and ulcer disease during our lifetime and this will increase patients taking corticosteroids.57 T h e role of with age.” These data need to be considered when addressing the epidemiology of NSAID g a ~ t r o p a t h y . ~ ~ corticosteroids as a risk factor was recently confirmed by Piper et Until recently there has been little Despite this, there is incontrovertible evidence that use evidence that differentiates one NSAID from another of NSAIDs is associated with the development of with respect to inducing serious upper G I side gastric ulcers, with ulcer complications (haemorrhage effect^.^' Studies which purport to show that certain and perforation) and with ulcer death (Table l).45-52 NSAIDs are more or less gastrotoxic need to be interThe best estimate of absolute risks of severe upper GI preted carefully because equivalent anti-inflammatory adverse effect seems to be approximately two cases of XONSTEROIDAI AX?’I-Ih’l.I.AMMATORY DRIJGS AND GI
Aust NZ J Med 1992; 22
687
doses are not always employed. Two recent studies, however, have demonstrated significant differences between N S A I D s in their association with hospitalisationi6 or overall t o x i ~ i t y . ' ~In summary, although the absolute risk of a significant gastrointestinal adverse reaction with NSAIDs is not high for the individual patient, the multiple effect of exposure and relative risk translate into NSAIDs being the likely direct cause between 20% and 30% of all cases of peptic ulcer complications in countries with high usage such as Australia.6o
had healed ulcers after eight weeks, significantly higher than those given ranitidine, 300 mg daily, in a comparator arm of the same study. However, the NSAID treated patients were a fairly small subgroup imbedded in a larger trial of idiopathic ulcers. Further studies are needed, with larger patients numbers, to compare omeprazole directly with histamine H, blockers (perhaps at increased dosage) and prostaglandins to answer the question of how NSAID induced ulcers should best be healed in patients who cannot easily have their NSAIDs ceased.
ULCER TREATMENT There is no doubt that gastric and duodenal ulcers in patients with rheumatoid and osteoarthritis heal despite continuing NSAID therapy. However, they do seem to take longer to do so. For NSAID associated gastric ulcer, the pooled healing rate from several studies with histamine H,-receptor antagonists averaged 67% at eight weeks.6i-66Many of these studies were flawed by small patient numbers and lack of controls or blinding. By comparison, eight week healing rates of close to 90% have been achieved in most studies with H, blockers for idiopathic More direct evidence that NSAIDs delay the healing of both gastric and duodenal ulcers comes from a recent open study in which 170 patients were randomised to continue or to stop the NSAIDs, while undergoing treatment with ranitidine. After four weeks, about 20% more patients with either ulcer type had healed if the NSAIDs were stopped rather than continued. 66 The substituted aluminium-sucrose compound (sucralfate) is often prescribed for NSAID ulceration. There is little evidence for its efficacy - a modest reduction in gastric lesion scores in one study with unbalanced groups at entry6R - despite some promising short term results showing protection against NSAID induced bleeding.69Synthetic prostaglandins accelerate NSAID ulcer healing, compared with placebo, despite continuation of NSAIDs. However, the mean eight week extrapolated healing from four studies was still only much the same as the pooled mean from the studies with H, blockers referred to above. These prostaglandin studies were all placebo controlled and blinded. T h e proportion of patients who healed by eight weeks in the placebo arm was only 23% (extrapolated), compared with expected values of about 50% for idiopathic gastric again highlighting the slow healing of NSAID ulcers when the NSAIDs are not ceased. Only one study has been performed so far, using the acid pump inhibitor omeprazole, in patients with gastric ulcer who continue to take NSAIDS.~'In the group given omeprazole, 40 mg daily, 95% of patients 688 Aust NZ J Med 1992; 22
PROPHYLAXIS Preventive treatment to reduce the risk of developing NSAID associated ulceration is possible but the choice of agent is a little complex. The data at present suggest that H, blockers are effective for preventing duodenal ulceration in patients taking NSAIDS.'"'~ However, they have been found wanting so far for gastric ulceration despite some promising evidence for reduction in the acute intra-mucosal haemorrhages caused by N S A I D S . ~ ~ Since most studies have found an increased relative risk for gastric ulcer but not duodenal ulcer in NSAID takers, it is possible that H, blockers are mainly protective against development or relapse of incidental H. pylori induced duodenal ulcers and that the role of NSAIDs is of secondary importance in such patients. Prostaglandins also have been shown to protect against duodenal ulcer in patients taking diclofenac, although this has been reported so far only in interim or abstract form.7RSeveral studies have demonstrated efficacy of the PGEl-analogue misoprostol for preventing NSAID associated gastric ulcers. These have ranged in duration from three to 12 months and have demonstrated reductions in ulcer incidence ranging between 60% and 90% compared with or with sucralfate.*' Acid pump blockers have not yet been studied for prophylaxis of NSAID ulcers, although two large studies are about to commence. None of the prophylaxis studies have been able to demonstrate a reduction in the frequency of ulcer complications in patients receiving NSAID but to do so would be an enormous task for a controlled trial in view of the low frequency (2% of ulcers) of such complications. However, it is likely, although admittedly unproven, that any intervention that lowers the incidence of ulcers will also lower the rate of ulcer complications. Who should receive prophylactic therapy for peptic ulceration is currently a hotly debated question. Cost benefit issues need to be addressed as well as the question of which subgroups of patients are most at risk. PRACTICAL ISSUES As mentioned previously, NSAIDs are very commonly prescribed and are often given for degenerative BROOKS A M I YEOMANS
musculoskeletal complaints rather than for inflammatory joint disease. In these conditions, however, pain relief does seem to be associated with improved quality of life.” T h e frequency of dyspepsia in these patients is high but is not predictive of the type nor indeed the presence of any gastrointestinal pathology. Although it has been suggested that patients with RA might be at particular risk, the data are not persuasive. Any increase could represent confounding by general disease severity due to poor overall health or coincident rasculitis and not any inherent liability to damage. 4 s Fries et al. have suggested, there are some 2,600 deaths each year in the USA and 20,000 hospitalisations with NSAID associated adverse G I events in patients with RA. These figures are comparable with those suggested for the United K i n g d ~ m . ~83’ T h e first thing for the doctor to ask is, does the patient require an NSAID? If the problem is non-inflammatory (osteoarthritis), then use of pure analgesic agents such as paracetamol, physical treatments and possibly the use of percutaneous N S A I D S ~should ~ be considered. Most NSAIDs are now given as enteric coated or slow release preparations although, as mentioned previously, this does not totally prevent adverse gastric events. Other factors which might promote ulceration, such as smoking and alcohol consumption, should be reduced to a minimum although studies suggesting that these strategies will influence NSAID induced peptic ulceration are lacking. At present there are no convincing data to support primary p r ~ p h y l a x i s There . ~ ~ have been some suggestions that routine protaglandin prophylaxis might be cost effe~tive~’.~’ but these have used values from the optimistic end of the current efficacy data” and others have disagreed.x8 If prophylactic treatment is not to be recommended at present for all, which high risk subgroups might be considered? Information is not particularly good to answer this, but a reasonable approach might be as followsPatients who have a past history of gastric ulceration or who are elderly ( > 60 years), have severe RA or are receiving concomitant corticosteroid therapy should be considered for prophylaxis with a prostaglandin (at present misoprostol). Those having a previous history of duodenal ulcer should be placed on maintenance therapy with an H,-receptor antagonist. Just how long these drugs should be continued is not known but is the subject of considerable debate in the literature.” Patients should be followed carefully to monitor for anaemia or ulcer symptoms that might suggest the need for more intensive treatment or withdrawal of NSAIDs. T h e newer antiulcer preparations such as prostaglandin analogues have great potential in the management of complicated rheumatic disease but do NONSTEROIDAL ANTI-INFLAMMATORY DRUGS AND GI
require proper economic evaluation before definitive guidelines for use can be made. m Date of submission: 13 February 1992
References 1. Clive Dhi, Stoff JS. Renal syndromes associated with nonsteroidal anti-inflammatory drugs. N Engl J Med 1984; 310: 563-72. 2. Gibson T. Non-steroidal anti-inflammatory drugs - another look. Br J Rheumatol 1989; 27: 87-90. 3. Paulus HE. FDA Arthritis Advisory Committee Meeting: Post
4.
5.
6. 7. 8.
marketing surveillance of non-steroidal anti-inflammatory drugs. Arthritis Rheum 1985; 28: 11168-9. Roth SH, Bennett RE. NSAIIIs and gastropathy: a rheumatologist’s review. J Rheumatol 1988; 15: 912-8. Henry DA, Johnston A, Dobson A, Duggan J. Fatal peptic ulcer complications and the use of- non-steroidal anti-inflammatory drugs, aspirin and corticosteroids. Br Med J 1987; 295: 1227-9. Gabriel SE, Bombardier C. NSAID induced ulcers. An emerging epidemic. J Rheurnatol 1990; 17: 1-4. Baum C, Kennedy DI,, Forbes LIB. Utilization of non-steroidal anti-inflammatory drugs. Arthritis Rheum 1985; 28: 686-92. Weber JCP. Epidemiology of adverse reactions to non-steroidal anti-inflammatory drugs. In: Rainsford KD, Velo G P (Eds.). Advances in inflammation, Vol. 6. New York Raven Press,
1984; 1-7. 9. Sol1 AH. Pathogenesis of peptic ulcer and implications for therapy. N Engl J Med 1990; 322: 909-16. 10. Yeomans ND, St John DJB, de Boer WGRM. Regeneration
ofthe gastric mucosa after aspirin-induced injury in the rat. Am
J Dig Dis 1973; 18: 773-80. 11. Bjarnason I, Zanelli G, Smith T et al. Non-steroidal antiinflammatory drug induced intestinal inflammation in humans. Gastroenterology 1987; 93: 480-9. 2. Hingson DJ, Ito S. Effect of aspirin and related compounds on the tine structure of mouse gastric mucosa. Gastroenterology 1971; 61: 156-7. 3. Hoae GB, Champion GD, Corrigan AB et al. The effects of
enteric coating of aspirin tablets on occult gastrointestinal blood loss. Aust NZ J Med 1977; 7: 600-4. 1. Skeljo MV, Giraud AS, Yeomans ND. Mechanisms of indomethacin induced gastric damage. Clin Exper Pharmacol Physiol 1990; (Suppl 16): 171. 15. Graham DY, Smith JC, Holmes GI et al. Non-steroidal antiinflammatory effect of sulindac sulfoxide and sulphide on gastric mucosa. Clin Pharmacol Ther 1988; 38: 65-70. 16. Carson JL, Strom BL, Morse MI. et al. The relative gastrointestinaltoxicity of non-steroidal anti-inflammatorydrugs. Arch Intern Med 1987; 147: 1054-9. 7. Whittle BJR, Vane JR. Prostanoids as regulators of gastrointestinal function. In: Johnson I,R (Ed.). Physiology of gastrointestinal tract, 2nd ed. New York: Raven Press, 1987; 143-9. 8. Lichtenberger ISM, Richards JR, Hills BA. Effects of 16,
16-dimethyl prostaglandin E, on surface hydrophobicity of aspirin-treated canine gastric mucosa. Gastroenterology 1985; 88: 308-14. 9. Hogan DI,,
Eisenberg JI.. Gastroduodenal bicarbonate production. Adv Intern hled 1988; 33: 385-408. 20. Redfern JS, Feldnian M. Role of endogenous prostaglandins in preventing gastrointestinal ulceration: Induction of ulcers by antibodies to prostaglandins. Gastroenterology 1989; 96: (suppl): 596-605. 21. Whittle BJR. Mechanisms underlying gastric mucosal damage
by indomethacin and bile salts, and the action of prostaglandins. Br J Pharmacol 1977; 60: 455-60. Aust NZ J Med 1992; 22
689
22. Satoh H, Guth PH, Grossman MI. Role of food in gastrointestinal ulceration produced by indomethacin in the rat. Gastroenterology 1982; 83: 210-5. 23. Ligumsky M ct al. Aspirin can inhibit gastric mucosal cyclooxygenase without causing lesions in the rat. Gastroenterology 1983; 84: 756-61. 24. Pihan G, Szabo S. Dimethyl PGG2 and thiosulphate increase gastric mucosal penetration of hypertonic NaCl and increase net transmucosal water flux. Gastroenterology 1988; 94: A354. 25. Parkes DA. Oxygen radicals: mediarors of gastrointestinal pathophysiology. Gut 1989; 30: 293-8. 26. Tarnawski A er al. Does indomethacin affect quality of experimental gastric ulcer healing? A quantitative histologic and ultrastructural analysis. Gastroenterology 1990; 98: A1 36. 27. Okabe S, Ogihara Y. Role of granulation tissue and angiogenesis in healing of chronic gastric ulcers in rats. In: Mechanisms of injury, protection and repair of the upper gastrointestinal tract. World Congress of Gastroenterology Satellite Symposium, Cairns, Qld 1990; 53. 28. Laine I., Marin-Sorensen Md, Weinstein WM. Is Campylobacter pylori (CP) prevalence lower in gastric ulcers (GUS)because of non-steroidal anti-inflammatory drugs (NSAID) use? A prospective evaluation. Gastroenterology 1989; 96: (Suppl 11): 282. 29. Graham DY, Lidsky MD, Cox AM et al. I.ong-term nonsteroidal anti-inflammatory drug use and lielicobacrer pylori infection. Gastroenterology 1991; 100: 1653-57. 30. Sol1 AH, Weinstein WM, Kurata J, M d a r t h y D. Non-steroidal anti-inflammatory drugs and peptic ulcer disease. Ann Intern Med 1991; 114: 307-19. 31. Baskin VI", er al. Aspirin-induced ultrastructural change in human gastric mucosa: correlation with potential difference. Ann Intern Med 1976; 85: 299-303. 32. Gilbert DA et al. Prevention of acute aspirin-induced gastric mucosal injury by 15-R-15 methyl prostaglandin E,: an endoscopic study. Gastroenterology 1984; 86: 339-45. 33. Graham DY, Smith JL, Spjut HJ, Torres E. Gastric adaptation: studies in humans during continuous aspirin administration. Gastroenterology 1988; 95: 327-33. 34. Graham DY, Smith JL. Aspirin and the stomach. Ann Intern Med 1986; 104: 390-8 35. Silvoso GR, Ivey KJ, Butt J H tf al. Incidence of gastric lesions in patients with rheumatoid disease on chronic aspirin therapy. Ann Intern Med 1979; 91: 517-20. 36. St John DJB, Yeomans NI), McDermott FT ef al. Adaptation of the gastric mucosa to repeated administration of aspirin in the rat. Am J Dig Dis 1973; 18: 881-6. 37. Skeljo MY, Giraud AS, Yeomans N D . Adaptation of rat gastric mucoa to repeat NSAID dosage varies with NSAID and is due to less deeper injury. Gastroenterology 1991; 100: A164. 38. Carson JL, Strom BL, Soper KA er a!. The association of nonsteroidal anti-inflammatory drugs with upper gastrointestinal bleeding. Arch Intern Med 1987; 147: 85-8. 39. Griffin MR, Ray WA, Schaffner W. Non-steroidal antiinflammatory drug use and death from peptic ulcer in elderly persons. Ann Intern Med 1988; 109: 359-63. 40. Muller P et al. Endoscopic evaluation of gastroduodenal tolerability of low-dose aspirin therapy. Gastroenterology 1990; 98: A92. 41. Leonards JR, Levy G. Gastrointestinal blood loss from aspirin and sodium salicylate tablets in man. Clin Pharm Thera 1972; 14: 62-6. 42. Dooley CP, Cohen H, Fitzgihbon PI. er al. Prevalence of Ifelicobactor pylori infection and histologic gastritis in asymptomatic persons. N Engl J Med 1989; 321: 1562-6. 43. Kurata JH. Ulcer epidemiology: an overview and proposed research framework. Gastroenterology 1989; 96: 569-80. 44. Hadler NM. There's the forest: the object lesson NSAID gastropathy. J Rheumatol 1990; 17: 280-2.
690
Aust N Z J &led 1992; 22
45. Ihggan JM, Dobson AJ, Johnson H, Fahey P. Peptic ulcer and non-steroidal anti-inflammatory agents. Gut 1986; 27: 929-33. 46. McIntosh JH,Blyth K, Piper DU'. Environmental factors in aetiology of chronic gastric ulcer: a case control study of exposure variables before the first symptoms. Gut 1988; 26: 789-98. 47. Somerville K, Faulkner G, Langman MJS. Non-steroidal antiinflammatory drugs and bleeding peptic ulcer. Lancet 1986; ii: 462-4. 48. Collier I) St J, Pain JA. Non-steroidal anti-inflammatory drugs and peptic ulcer perforation. Gut 1985; 26: 359-63. 49. Smedley FH, Taube M, Leach R, Wastell C. Non-steroidal antiinflammatory drugs: retrospective study of bleeding and perforated peptic ulcers (ilbst). Gut 1988; 29: A1443. 50. Hawkey CJ. Non-steroidal anti-inflammatory drugs and peptic ulcer. Br Med J 1990; 300: 2i8-84. 51. Jick SS, Perera DR, Walker AM, Jick H.Non-steroidal antiinflammatory drugs and hospital admission for perforated peptic ulcer. Lancet 1987; ii: 380-2. 52. Quader K, Logan RFA. Peptic ulcer ( I T ) deaths; how many occur at home or after non-steroidal anti-inflammatory drug (NSAID) prescribing? Gut 1988; 29: A1443. 53. Langman MJ. Epidemiological evidence on the association between peptic ulceration and anti-inflammatory drug use. Gastro Enterol 1989; 96 (Suppl): 640-6. 54. Beard K, Walker :LU, Perera DR, Jick H. Non-steroidal antiintlammatory drug and hospitalisation for gastroesophageal bleeding in the elderly. Arch Intern Med 1987; 147: 1621-3. 55. I'angman MJS, Brooks P, Hawkey CJ, Silverstein F, Yeomans N. Non-steroidal anti-inflammatory drug associated ulcer: epidemiology, causation and treatment. J Gastroenterol Hepatol 1991; 6: 442-9. 56. Laporte JR, Came X, Vidal X, Moreno V, Juan J. LJpper gastrointestinal bleeding in relation to previous use of analgesics and non-steroidal anti-inflammatory drugs. Lancet 1991; 337: 85-9. 57. Fries JF, Miller SR, Spitz PW, Williams CA, Hubert HB, Bloch DA. Toward an epidemiology of gastropathy associated with non-steroidal anti-inflammatorydrug use. Gastroenterology 1989; 96: 647-55. 58. Piper JM, Ray Wh, Daugherty JR, Griffin MK. Corticosteroid use and peptic ulcer disease: role of non-steroidal antiinflammatory drugs. Ann Intern Med 1991; 114: 735-40. 59. lries JF, Williams CA, Bloch DA. The relative toxicity ofnonsteroidal anti-inflammatory drugs. Arthritis Rheum 1991; 34: 1353-60. 60. Henry DA, Johnson A, Dobson A, Iluggan J. Fatal peptic ulcer complications and the use of' non-steroidal anti-inflammatory drugs, aspirin and corticosteroids. Br Med J 1987; 295: 1227-9. 61. Bijlsma JWJ. Treatment of MAID-induced gastrointestinal lesions with cimetidine: an international multicentre collaborative study. Aliment Pharmacol Ther 1982; 2 (Suppl 1): 89-96. 62. Croker JR, Cotton PR, Boyle AC, Kinsella P. Cimetidine for peptic ulcer in patients with arthritis. Ann Rheum Dis 1980; 39: 275-8. 63. lhvies J, Collins AJ, Dixon A StJ. T h e influence of cimetidine on peptic ulcer in patients with arthritis taking anti-inflammatory drugs. Br J Rheumatol 1986; 25: 54-8. 64. Farah D, Sturrock RI), Russell RI. Peptic ulcer in rheumatoid arthritis. Ann Rheum Dis 1988; 47: 478-80. 65. Walan A, Rader JP, Classen M, Lamaers CB, Piper DW. Effect of omeprazole and ranitidine on ulcer healing and relapse rates in patients with benign gastric ulcer. N Engl J Med 1989; 320: 69-75. 66. 1.ancaster-Smith MJ, Jaderberg ME, Jackson DA. Ranitidine in the treatment of NSAID-associated gastric and duodenal ulcers. Gut 1991; 32: 252-5. 67. Isenberg JI, Peterson WL, Elashoff JD er al. Healing of benign gastric ulcer with low-dose antacid or cimetidine. A double-blind, randomized, placebo-controlled trial. N Engl J Med 1983; 308: 1319-24. RROOKS AND YEOiMANS
68. Caldwell JR, Roth SH, Wu WC p r a/. Sucralfate treatment of non-steroidal anti-inflammatory drug-induced gastrointestinal symptoms and mucosal damage. Am J Med 1987; 8 3 (Suppl 3B): 74-82. 69. Malchow-Moller A, Ranlov PJ. Does sucralfate reduce acetylsalicylic acid-induced gastric mucosal bleeding? Scdnd J Gastroenterol 1987; 22: 550-2. 70. Rorh S, Agrawal N, Mahowald M ~f a/. Misoprostol heals gastroduodenal injury in patients with rheumatoid arthritis receiving aspirin. Arch Intern Med 1989; 149: 775-9. 71. Paladini G, Fabiani MG, Tosato F, Maggiolo F. Prophylactic and therapeutic role of rioprostil in NSAIDs induced gastroduodenal lesions. Scand J Gastroenterol 1989; 224 (Suppl 164): 242-6. 72. Sontag SJ, Schnell TG, Mak E el a/. Enprosti1 heals NSAID induced gastric ulcers. Gastroenterology 1990; 98: A1 29. 73. Euler AR, Safdi M, Rao j er QI. A report of three multiclinic trials evaluating arhaprostil in arthritic patients with ASANSAID gastric mucosal damage. Gastroenterology 1990; 98: 1549-57. 74. Shearman DJC, Finlayson NDC. Diseases of the gastrointestinal tract and liver, 2nd ed. Edinburgh: Churchill Livingstune 1989. 75. Ehsanalluh RSB, Page MC, Tildesley G er al. Prevention of gastroduodenal damage induced by non-steroidal anti-inflammatory drugs: controlled trial ofranitidine. Br Med J 1988; 297: 1017-21. 76. Kobinson MG, GrifFin JW Jr, Rowers J, Kogan Fj er al. Effect of ranitidine on gastroduodenal mucosal damage induced by nonsteroidal anti-inflammatory drugs. Dig Dis Sci 1989; 34: 424-8. 77. Berkowitz JM, Rogenes PR, Sharp J?',Warner CW. Ranitidine protects against gastroduodenal mucosal damage associated with chronic aspirin therapy. Arch Intern Med 1987; 147: 2137-9. 78. Geis GS, Stead H, Wallemark CB, Nicholson PA. Prevalence of mucosal lesions in the stomach and duodenum due to chronic use of NSAID in patients with rheumatoid arthritis or osteoarthritis and interim report on prevention of diclofenac associated lesions. J Rheumatol Suppl 1991; 28: 11-4.
NONSTEROIDAL ANTI-1NFI.AILIMATORY DRUGS AND GI
79. Graham IIU, Agrawal NM, Roth SH. Prevention of NSAIDinduced gastric ulcer with rnisoprostol multicentre double blind placebo controlled trial. Lancet 1988; ii: 1277-80. 80. Elliott SL, Yeonians ND, Buchanan RRC er 11. Long term effects of misoprostol on gastropathy induced by non-steroidal antiinflammatory drugs (NSAIIh). Gastroenterology 1990; 8: S59. 81. Agrawal N, Stromatt S, Brown J. Comparative study of misoprostol and sucralfate in rhe prevention of NSAID induced gastric ulcers. Clin Exp Rheuniatol 1990; 8: 59. 82. Parr G, Darekar B, Fletcher A, Rulpitt CJ. Joint pain and quality of life: results of a randomised trial. Br J Clin Pharmacol 1989; 27: 235-42. 83. Armstrong CP, Blower AL. Non-steroidal anti-inflammatory drugs and life threatening complications of peptic ulceration. Gut 1987; 28: 527-32. 84. Editorial. Topical NSAIDs: a gimmick or a godsend. Lancet 1989; 779-80. 85. Gabriel SE. Is misoprostol prophylaxis indicated for NSAID induced adverse gastrointestinal events? An epidemiologic opinion. J Rheumatol 1991; 18: 958-61. 86. Hillman AI,, Bloom BS. Economic effects of prophylactic use of misoprostol to prevent gastric ulcer in patients taking nonsteroidal anti-inflammatory drugs. Arch Intern Med 1989; 149: 2061-5. 87. Knill-Jones R, Drumniond M, Kohli H, Davies I,. Economic evaluation of gastric ulcer prophylaxis in patients with arthritis receiving non-steroidal anti-inflammatory drugs. Postgrad Med J 1990; 66: 639-46. 88. Edelson JT, Tosreson ANA, Sax P. Cost effectiveness of misoprostol for prophylaxis against non-aeroidal anti-inflammatory druginduced gastrointestinal bleeding. JAMA 1990; 264: 41-7. 89. Barrison I. Prophylaxis against non-steroidal induced upper gastrointestinal side effects. Ann Rheum Dis 1991; 5 0 207-9.
Aust NZ J Med 1992; 22
691
Nonsteroidal anti-inflammatory drug gastropathy is it preventable?
-
P. M. Brooks Professor of Medicinc, Department of Medicine, I!niversiry of New South Wales, St Vincent’s IIospital, Sydney, NSW
N. D. Yeomans Professor of hledicine, Department of Medicine, University of Melbourne, Western Hospital, Melbourne, Vic
INTRODUCTION The non-steroidal anti-inflammatory drugs (NSAIDs) make up a group of the most commonly prescribed pharmaceuticals.‘ It is estimated that more than 30 million people throughout the world consume an NSAID each day.* T h e number of NSAID prescriptions written each year in the USA doubled between 1984 and 19883,4while in Australia the prevalence of NSAID use in elderly subjects has been estimated to be in excess of 20Y0.~T o a large extent this is a reflection of the high prevalence of rheumatic diseases in the community but an annual growth rate of 6% in NSAID prescriptions as seen in Canada during 1989 is unlikely to be due to increased disease incidence alone.’ As the population ages it is at risk from painful musculoskeletal diseases such as osteoarthritis, for which NSAIDs may not be the most appropriate treatment, and yet it seems that it is for these primarily non-inflammatory forms of arthritis that the majority of NSAIDs are prescribed.’ NSAIDs are associated with a wide spectrum of side effects including those involving renal, hepatic and haematological hnction and skin rashes. It is, however, the gastrointestinal (GI) side effects of NSAIDs which are most common and serious and make NSAIDs the most frequent cause of drug side effects reported to the Committee on Safety of Medicines in Britain each year.8 It is because of this that it is important to try to ascertain the true relationship between NSAID use and the clinically important gastrointestinal side effects and to develop strategies which might diminish this problem. In this review we will discuss briefly the pathophysiology of so-called ‘NSAID gastropathy’,
review the epidemiology of NSAID associated G I side effects and examine the clinical trial evidence that drugs might influence the ulcer healing rate and perhaps even prevent their development.
PATHOPHYSIOLOGY OF NSAID GASTROPATHY Mucosal damage in the gastrointestinal tract is the product of a fine balance between acid production and mucosal events. Surface epithelial cells secrete mucus and bicarbonateg and have the ability to rapidly repair injury.l0 These properties are essential for maintaining mucosal integrity in the stomach. Environmental factors such as smoking and stress, genetic predisposition, the bacterium Helicobacter pylori and NSAIDs (Figure 1) all play a role in predisposing to peptic ulceration. The epithelial cells can be damaged by acid, pepsin, refluxed bile acids, NSAIDs or the products of H. pylori. NSAIDs produce a wide spectrum of injury to the gastrointestinal tract ranging from petechial haemorrhages and erosions to ulcers. It is clear that the whole of the gastrointestinal tract is at risk from NSAID associated damage” with small and large bowel mucosal ulceration noted. NSAIDs probably have both a local and a systemic influence on G I mucosal physiology. Aspirin produces much of its damage topically, after it has entered the gastric surface mucus cells from the lumen. Evidence for this includes the high speed with which aspirin causes surface injury12 and the partial protection against injury which is provided by enteric coating of the aspirin.l3 However, the non-aspirin NSAIDs seem to produce most of their injury systemically, after they have been
Covresponiicwe: Professor 1’. M. Brooks, Ikpartment of Medicine, I’nlversiry of New South Wales, St Vincent’s Hospital, Darlinghurst, M W 2010, Australia. XONS’IEROIDAI, A N ’ I ‘ I - I ~ ~ ~ L A . M M A I OIIKCGS RY AND (;I
Aust NZ J Med 1992; 22
685
-
Arachidonic acid
MUCUS& HC03
NSAlDs
block synthesis
ndins
Microcirculation Figure I: Possible sites of action for the ‘cytoprotective’ acrions of endogenous prostaglandins in the gastric mucosa. I’rosragkdndins stimulate secrerion of mucus and bicarbonate by the rpithelial cells and also appear to stabilise the microcirculation.
absorbed. For example, in rat^,'^ parenteral administration of several commonly used NSAIDs produces at least as much gastric damage as does oral administration. Furthermore, studies in humans have shown that pro-drugs such as sulindac, enterically coated NSAIDs and rectally administered NSAIDs are all still associated with gastric damage - both acute and chr~nic,’’,’~albeit less frequently than with oral administration. It is possible that much of the gastric injury produced by NSAIDs is a consequence of the inhibition of the production of prostaglandins by the mucosal cells. This is caused by a blockade of cyclooxygenase at the start of the prostanoid biosynthetic pathway which is an effect common to all NSAIDs and, therefore, all NSAIDs share the capacity to produce gastric injury. There is now mounting evidence for several protective effects of gastric prostaglandins. They stimulate mucus and bicarbonate secretion from various cells” and seem to also increase the capacity of the mucosal surface to repel luminal acid - by increasing its hydrophobicity or ‘non-wettability’.18 Conversely, bicarbonate secretion is inhibited by several NSAIDS.’~ Moreover, severe gastric ulcers can be produced in experimental animals by inducing or administering antibodies to prostaglandins but not by antibodies to inactive prostaglandin analogues.20 Inhibition of prostaglandin production might enhance subsceptibility to damage by bile acids, hypertonic saline and even f o ~ d . * Prostaglandins ~-~~ might also stimulate 686 Aust NZ J Med 1992; 22
serosal-mucosal water flux so histodiluting injurious
agent^.'^ Some of the injurious effects of NSAIDs might be independent of blockade of prostaglandins. For instance, there is evidence that NSAIDs might be involved in the production of oxygen-free radicals.25 They may also impair angiogenesis at the base and margins of the ulcer, leading to ulcers which are more fibrous and less contractile than ~ s u a l . ~ ~ . ~ ’ T h e interaction between H. pylori and NSAIIh may also be important. However, evidence that these are independent risk factors for ulcers is mounting. One study in patients presenting with gastric ulcer found I€. pylori colonisation in nine of ten of the patients who were not taking NSAIDs but in only five of 12 of those taking the drugs.28 In a much larger study, Graham, et aL2’ found that the frequency of H. pylori infection was independent of NSAID use and that NSAID induced erosions were no more frequent in those with H. pylori infection than those without. It is, therefore, possible that there are two types of NSAID induced ulcers - ulcers that occur de ~ ~ o vino histologically normal gastric niucosa and ulcers that occur as a result of NSAID a d m i n i ~ t r a t i o n . ~ ~ T h e sequence of gastric mucosal events following ingestion of soluble aspirin has been reasonably well documented. Changes in the epithelial cells have been noted within minutes of aspirin ingestion,?‘ and by two hours nearly all subjects given a single dose will have red mucosal lesions and/or erosions. The subepithelial haemorrhages come and go during therapy with RKOOKS AND YEOl\l..lNS
NSAIDs.”*However, erosions occasionally fail to heal and progress to form ulcers. These, in turn, will often heal s p o n t a n e o ~ s l y It . ~ is ~ ~not ~ ~ clear, ~ however, in what proportion of patients this sequence of events is followed. For example, not all patients taking NSAIDs chronically will have erosions or ulcers when endoscoped. Silvoso, er a/., ’’ have demonstrated that patients taking eight aspirin tablets a day for at least three months show gastric erosions in 40% and gastric ulcers in 16%. This suggests that a large proportion of patients have adapted to NSAID challenge as suggested by Graham, er al. In these studies Graham showed that decreasing damage was apparent with continued exposure to aspirin and that lesions healed more rapidly if given a longer period of adaptive exposure to a ~ p i r i n . ”Adaptation ~ was more rapid at lower aspirin doses (1.3 gm per day) than higher doses (2.4 gm per day). This adaptation took several weeks to develop, during which time there was a gradual return to an almost normal mucosal appearance. Similar adaptation, although with a more rapid onset, has been seen in experimental animal^.^^,^' These experimental data are paralleled by epidemiological evidence. Carson, et a/.”x showed that the risk of gastrointestinal bleeding increased to a maximum after four NSAID prescriptions and decreased to no measurable risk after ten prescriptions. Griffin, et have shown that the maximal risk of gastrointestinal bleeding occurs in the first months after an NSAID prescription. O n the other hand, other groups have found no evidence of adaptation in humans using endoscopic assessment or measurement of gastrointestinal blood loss.“o,“l The majority of these studies have been done in patients who do not have chronic rheumatoid arthritis or osteoarthritis. There are, however, no clear data from the literature to suggest that there are differences between rheumatoid and non-rheumatoid gastric mucosa.
TABLE 1 _.__-___
9540 Ulct:r occurence
Ref
Reldtive risk
con‘idence interval
_ _ _ ~ _ _ _-~_ Gastric
Dmdenal
45 46
5 47
’3166
46
1 1
0 4-3 7
47 47
28 27
1554 1358
48 49 51
73 29 16
44’1 8 ’174 0737
39 39 53
42 79 29
1990 3 7-16 8 1 4 - 63
14769
Bleediyj
yastric tluotJerldl
Per‘oration peptic
(Jlcer dedtr
gastric duodenal pevtic
_ ~ _ _ _ _
-~
NSAID induced reaction per 10,000 person months of prescription^.^^ Using the technique of automated record linkage which relates diagnostic category to drug use incorporating information contained in large data bases, relative risks for perforated ulcer or haemorrhage as a result of NSAID therapy have been calculated to be around 1.S4Prospective case control studies, which probably give a more reliable estimate, indicate a substantially higher relative risk of serious upper G I side effects of between three and seven.j5 5 6 Fries et ul.,” using multivariate analysis in 2,400 patients with rheumatoid arthritis (RA) followed prospectively for an average period of 3.5 years, have demonstrated a hazard ratio (approximating to relative risk) for hospitalisations due to GI adverse effects of seven times that for patients not on NSAIDs. These data have led Fries et al. to suggest that in the United States the syndrome of NSAID induced EPIDEMIOLOGY gastropathy can be considered to account for at least The first point to remember is that diseases of the 2,600 deaths per annum and 20,000 hospitalisations gastric mucosa are extremely prevalent, with approxieach year in patients with RA alone. T h e studies of mately 10% of healthy asymptomatic people having Fries et al. suggest that those patients particularly at mucosal lesions at any one time and as many as 1% risk for development of NSAID related gastropathy showing extension into the submucosa, i.e., ulcer.42 are the elderly, those with a previous history of peptic Approximately 10% of us will experience frank peptic ulceration, those with severe rheumatoid disease and ulcer disease during our lifetime and this will increase patients taking corticosteroids.57 T h e role of with age.” These data need to be considered when addressing the epidemiology of NSAID g a ~ t r o p a t h y . ~ ~ corticosteroids as a risk factor was recently confirmed by Piper et Until recently there has been little Despite this, there is incontrovertible evidence that use evidence that differentiates one NSAID from another of NSAIDs is associated with the development of with respect to inducing serious upper G I side gastric ulcers, with ulcer complications (haemorrhage effect^.^' Studies which purport to show that certain and perforation) and with ulcer death (Table l).45-52 NSAIDs are more or less gastrotoxic need to be interThe best estimate of absolute risks of severe upper GI preted carefully because equivalent anti-inflammatory adverse effect seems to be approximately two cases of XONSTEROIDAI AX?’I-Ih’l.I.AMMATORY DRIJGS AND GI
Aust NZ J Med 1992; 22
687
doses are not always employed. Two recent studies, however, have demonstrated significant differences between N S A I D s in their association with hospitalisationi6 or overall t o x i ~ i t y . ' ~In summary, although the absolute risk of a significant gastrointestinal adverse reaction with NSAIDs is not high for the individual patient, the multiple effect of exposure and relative risk translate into NSAIDs being the likely direct cause between 20% and 30% of all cases of peptic ulcer complications in countries with high usage such as Australia.6o
had healed ulcers after eight weeks, significantly higher than those given ranitidine, 300 mg daily, in a comparator arm of the same study. However, the NSAID treated patients were a fairly small subgroup imbedded in a larger trial of idiopathic ulcers. Further studies are needed, with larger patients numbers, to compare omeprazole directly with histamine H, blockers (perhaps at increased dosage) and prostaglandins to answer the question of how NSAID induced ulcers should best be healed in patients who cannot easily have their NSAIDs ceased.
ULCER TREATMENT There is no doubt that gastric and duodenal ulcers in patients with rheumatoid and osteoarthritis heal despite continuing NSAID therapy. However, they do seem to take longer to do so. For NSAID associated gastric ulcer, the pooled healing rate from several studies with histamine H,-receptor antagonists averaged 67% at eight weeks.6i-66Many of these studies were flawed by small patient numbers and lack of controls or blinding. By comparison, eight week healing rates of close to 90% have been achieved in most studies with H, blockers for idiopathic More direct evidence that NSAIDs delay the healing of both gastric and duodenal ulcers comes from a recent open study in which 170 patients were randomised to continue or to stop the NSAIDs, while undergoing treatment with ranitidine. After four weeks, about 20% more patients with either ulcer type had healed if the NSAIDs were stopped rather than continued. 66 The substituted aluminium-sucrose compound (sucralfate) is often prescribed for NSAID ulceration. There is little evidence for its efficacy - a modest reduction in gastric lesion scores in one study with unbalanced groups at entry6R - despite some promising short term results showing protection against NSAID induced bleeding.69Synthetic prostaglandins accelerate NSAID ulcer healing, compared with placebo, despite continuation of NSAIDs. However, the mean eight week extrapolated healing from four studies was still only much the same as the pooled mean from the studies with H, blockers referred to above. These prostaglandin studies were all placebo controlled and blinded. T h e proportion of patients who healed by eight weeks in the placebo arm was only 23% (extrapolated), compared with expected values of about 50% for idiopathic gastric again highlighting the slow healing of NSAID ulcers when the NSAIDs are not ceased. Only one study has been performed so far, using the acid pump inhibitor omeprazole, in patients with gastric ulcer who continue to take NSAIDS.~'In the group given omeprazole, 40 mg daily, 95% of patients 688 Aust NZ J Med 1992; 22
PROPHYLAXIS Preventive treatment to reduce the risk of developing NSAID associated ulceration is possible but the choice of agent is a little complex. The data at present suggest that H, blockers are effective for preventing duodenal ulceration in patients taking NSAIDS.'"'~ However, they have been found wanting so far for gastric ulceration despite some promising evidence for reduction in the acute intra-mucosal haemorrhages caused by N S A I D S . ~ ~ Since most studies have found an increased relative risk for gastric ulcer but not duodenal ulcer in NSAID takers, it is possible that H, blockers are mainly protective against development or relapse of incidental H. pylori induced duodenal ulcers and that the role of NSAIDs is of secondary importance in such patients. Prostaglandins also have been shown to protect against duodenal ulcer in patients taking diclofenac, although this has been reported so far only in interim or abstract form.7RSeveral studies have demonstrated efficacy of the PGEl-analogue misoprostol for preventing NSAID associated gastric ulcers. These have ranged in duration from three to 12 months and have demonstrated reductions in ulcer incidence ranging between 60% and 90% compared with or with sucralfate.*' Acid pump blockers have not yet been studied for prophylaxis of NSAID ulcers, although two large studies are about to commence. None of the prophylaxis studies have been able to demonstrate a reduction in the frequency of ulcer complications in patients receiving NSAID but to do so would be an enormous task for a controlled trial in view of the low frequency (2% of ulcers) of such complications. However, it is likely, although admittedly unproven, that any intervention that lowers the incidence of ulcers will also lower the rate of ulcer complications. Who should receive prophylactic therapy for peptic ulceration is currently a hotly debated question. Cost benefit issues need to be addressed as well as the question of which subgroups of patients are most at risk. PRACTICAL ISSUES As mentioned previously, NSAIDs are very commonly prescribed and are often given for degenerative BROOKS A M I YEOMANS
musculoskeletal complaints rather than for inflammatory joint disease. In these conditions, however, pain relief does seem to be associated with improved quality of life.” T h e frequency of dyspepsia in these patients is high but is not predictive of the type nor indeed the presence of any gastrointestinal pathology. Although it has been suggested that patients with RA might be at particular risk, the data are not persuasive. Any increase could represent confounding by general disease severity due to poor overall health or coincident rasculitis and not any inherent liability to damage. 4 s Fries et al. have suggested, there are some 2,600 deaths each year in the USA and 20,000 hospitalisations with NSAID associated adverse G I events in patients with RA. These figures are comparable with those suggested for the United K i n g d ~ m . ~83’ T h e first thing for the doctor to ask is, does the patient require an NSAID? If the problem is non-inflammatory (osteoarthritis), then use of pure analgesic agents such as paracetamol, physical treatments and possibly the use of percutaneous N S A I D S ~should ~ be considered. Most NSAIDs are now given as enteric coated or slow release preparations although, as mentioned previously, this does not totally prevent adverse gastric events. Other factors which might promote ulceration, such as smoking and alcohol consumption, should be reduced to a minimum although studies suggesting that these strategies will influence NSAID induced peptic ulceration are lacking. At present there are no convincing data to support primary p r ~ p h y l a x i s There . ~ ~ have been some suggestions that routine protaglandin prophylaxis might be cost effe~tive~’.~’ but these have used values from the optimistic end of the current efficacy data” and others have disagreed.x8 If prophylactic treatment is not to be recommended at present for all, which high risk subgroups might be considered? Information is not particularly good to answer this, but a reasonable approach might be as followsPatients who have a past history of gastric ulceration or who are elderly ( > 60 years), have severe RA or are receiving concomitant corticosteroid therapy should be considered for prophylaxis with a prostaglandin (at present misoprostol). Those having a previous history of duodenal ulcer should be placed on maintenance therapy with an H,-receptor antagonist. Just how long these drugs should be continued is not known but is the subject of considerable debate in the literature.” Patients should be followed carefully to monitor for anaemia or ulcer symptoms that might suggest the need for more intensive treatment or withdrawal of NSAIDs. T h e newer antiulcer preparations such as prostaglandin analogues have great potential in the management of complicated rheumatic disease but do NONSTEROIDAL ANTI-INFLAMMATORY DRUGS AND GI
require proper economic evaluation before definitive guidelines for use can be made. m Date of submission: 13 February 1992
References 1. Clive Dhi, Stoff JS. Renal syndromes associated with nonsteroidal anti-inflammatory drugs. N Engl J Med 1984; 310: 563-72. 2. Gibson T. Non-steroidal anti-inflammatory drugs - another look. Br J Rheumatol 1989; 27: 87-90. 3. Paulus HE. FDA Arthritis Advisory Committee Meeting: Post
4.
5.
6. 7. 8.
marketing surveillance of non-steroidal anti-inflammatory drugs. Arthritis Rheum 1985; 28: 11168-9. Roth SH, Bennett RE. NSAIIIs and gastropathy: a rheumatologist’s review. J Rheumatol 1988; 15: 912-8. Henry DA, Johnston A, Dobson A, Duggan J. Fatal peptic ulcer complications and the use of- non-steroidal anti-inflammatory drugs, aspirin and corticosteroids. Br Med J 1987; 295: 1227-9. Gabriel SE, Bombardier C. NSAID induced ulcers. An emerging epidemic. J Rheurnatol 1990; 17: 1-4. Baum C, Kennedy DI,, Forbes LIB. Utilization of non-steroidal anti-inflammatory drugs. Arthritis Rheum 1985; 28: 686-92. Weber JCP. Epidemiology of adverse reactions to non-steroidal anti-inflammatory drugs. In: Rainsford KD, Velo G P (Eds.). Advances in inflammation, Vol. 6. New York Raven Press,
1984; 1-7. 9. Sol1 AH. Pathogenesis of peptic ulcer and implications for therapy. N Engl J Med 1990; 322: 909-16. 10. Yeomans ND, St John DJB, de Boer WGRM. Regeneration
ofthe gastric mucosa after aspirin-induced injury in the rat. Am
J Dig Dis 1973; 18: 773-80. 11. Bjarnason I, Zanelli G, Smith T et al. Non-steroidal antiinflammatory drug induced intestinal inflammation in humans. Gastroenterology 1987; 93: 480-9. 2. Hingson DJ, Ito S. Effect of aspirin and related compounds on the tine structure of mouse gastric mucosa. Gastroenterology 1971; 61: 156-7. 3. Hoae GB, Champion GD, Corrigan AB et al. The effects of
enteric coating of aspirin tablets on occult gastrointestinal blood loss. Aust NZ J Med 1977; 7: 600-4. 1. Skeljo MV, Giraud AS, Yeomans ND. Mechanisms of indomethacin induced gastric damage. Clin Exper Pharmacol Physiol 1990; (Suppl 16): 171. 15. Graham DY, Smith JC, Holmes GI et al. Non-steroidal antiinflammatory effect of sulindac sulfoxide and sulphide on gastric mucosa. Clin Pharmacol Ther 1988; 38: 65-70. 16. Carson JL, Strom BL, Morse MI. et al. The relative gastrointestinaltoxicity of non-steroidal anti-inflammatorydrugs. Arch Intern Med 1987; 147: 1054-9. 7. Whittle BJR, Vane JR. Prostanoids as regulators of gastrointestinal function. In: Johnson I,R (Ed.). Physiology of gastrointestinal tract, 2nd ed. New York: Raven Press, 1987; 143-9. 8. Lichtenberger ISM, Richards JR, Hills BA. Effects of 16,
16-dimethyl prostaglandin E, on surface hydrophobicity of aspirin-treated canine gastric mucosa. Gastroenterology 1985; 88: 308-14. 9. Hogan DI,,
Eisenberg JI.. Gastroduodenal bicarbonate production. Adv Intern hled 1988; 33: 385-408. 20. Redfern JS, Feldnian M. Role of endogenous prostaglandins in preventing gastrointestinal ulceration: Induction of ulcers by antibodies to prostaglandins. Gastroenterology 1989; 96: (suppl): 596-605. 21. Whittle BJR. Mechanisms underlying gastric mucosal damage
by indomethacin and bile salts, and the action of prostaglandins. Br J Pharmacol 1977; 60: 455-60. Aust NZ J Med 1992; 22
689
22. Satoh H, Guth PH, Grossman MI. Role of food in gastrointestinal ulceration produced by indomethacin in the rat. Gastroenterology 1982; 83: 210-5. 23. Ligumsky M ct al. Aspirin can inhibit gastric mucosal cyclooxygenase without causing lesions in the rat. Gastroenterology 1983; 84: 756-61. 24. Pihan G, Szabo S. Dimethyl PGG2 and thiosulphate increase gastric mucosal penetration of hypertonic NaCl and increase net transmucosal water flux. Gastroenterology 1988; 94: A354. 25. Parkes DA. Oxygen radicals: mediarors of gastrointestinal pathophysiology. Gut 1989; 30: 293-8. 26. Tarnawski A er al. Does indomethacin affect quality of experimental gastric ulcer healing? A quantitative histologic and ultrastructural analysis. Gastroenterology 1990; 98: A1 36. 27. Okabe S, Ogihara Y. Role of granulation tissue and angiogenesis in healing of chronic gastric ulcers in rats. In: Mechanisms of injury, protection and repair of the upper gastrointestinal tract. World Congress of Gastroenterology Satellite Symposium, Cairns, Qld 1990; 53. 28. Laine I., Marin-Sorensen Md, Weinstein WM. Is Campylobacter pylori (CP) prevalence lower in gastric ulcers (GUS)because of non-steroidal anti-inflammatory drugs (NSAID) use? A prospective evaluation. Gastroenterology 1989; 96: (Suppl 11): 282. 29. Graham DY, Lidsky MD, Cox AM et al. I.ong-term nonsteroidal anti-inflammatory drug use and lielicobacrer pylori infection. Gastroenterology 1991; 100: 1653-57. 30. Sol1 AH, Weinstein WM, Kurata J, M d a r t h y D. Non-steroidal anti-inflammatory drugs and peptic ulcer disease. Ann Intern Med 1991; 114: 307-19. 31. Baskin VI", er al. Aspirin-induced ultrastructural change in human gastric mucosa: correlation with potential difference. Ann Intern Med 1976; 85: 299-303. 32. Gilbert DA et al. Prevention of acute aspirin-induced gastric mucosal injury by 15-R-15 methyl prostaglandin E,: an endoscopic study. Gastroenterology 1984; 86: 339-45. 33. Graham DY, Smith JL, Spjut HJ, Torres E. Gastric adaptation: studies in humans during continuous aspirin administration. Gastroenterology 1988; 95: 327-33. 34. Graham DY, Smith JL. Aspirin and the stomach. Ann Intern Med 1986; 104: 390-8 35. Silvoso GR, Ivey KJ, Butt J H tf al. Incidence of gastric lesions in patients with rheumatoid disease on chronic aspirin therapy. Ann Intern Med 1979; 91: 517-20. 36. St John DJB, Yeomans NI), McDermott FT ef al. Adaptation of the gastric mucosa to repeated administration of aspirin in the rat. Am J Dig Dis 1973; 18: 881-6. 37. Skeljo MY, Giraud AS, Yeomans N D . Adaptation of rat gastric mucoa to repeat NSAID dosage varies with NSAID and is due to less deeper injury. Gastroenterology 1991; 100: A164. 38. Carson JL, Strom BL, Soper KA er a!. The association of nonsteroidal anti-inflammatory drugs with upper gastrointestinal bleeding. Arch Intern Med 1987; 147: 85-8. 39. Griffin MR, Ray WA, Schaffner W. Non-steroidal antiinflammatory drug use and death from peptic ulcer in elderly persons. Ann Intern Med 1988; 109: 359-63. 40. Muller P et al. Endoscopic evaluation of gastroduodenal tolerability of low-dose aspirin therapy. Gastroenterology 1990; 98: A92. 41. Leonards JR, Levy G. Gastrointestinal blood loss from aspirin and sodium salicylate tablets in man. Clin Pharm Thera 1972; 14: 62-6. 42. Dooley CP, Cohen H, Fitzgihbon PI. er al. Prevalence of Ifelicobactor pylori infection and histologic gastritis in asymptomatic persons. N Engl J Med 1989; 321: 1562-6. 43. Kurata JH. Ulcer epidemiology: an overview and proposed research framework. Gastroenterology 1989; 96: 569-80. 44. Hadler NM. There's the forest: the object lesson NSAID gastropathy. J Rheumatol 1990; 17: 280-2.
690
Aust N Z J &led 1992; 22
45. Ihggan JM, Dobson AJ, Johnson H, Fahey P. Peptic ulcer and non-steroidal anti-inflammatory agents. Gut 1986; 27: 929-33. 46. McIntosh JH,Blyth K, Piper DU'. Environmental factors in aetiology of chronic gastric ulcer: a case control study of exposure variables before the first symptoms. Gut 1988; 26: 789-98. 47. Somerville K, Faulkner G, Langman MJS. Non-steroidal antiinflammatory drugs and bleeding peptic ulcer. Lancet 1986; ii: 462-4. 48. Collier I) St J, Pain JA. Non-steroidal anti-inflammatory drugs and peptic ulcer perforation. Gut 1985; 26: 359-63. 49. Smedley FH, Taube M, Leach R, Wastell C. Non-steroidal antiinflammatory drugs: retrospective study of bleeding and perforated peptic ulcers (ilbst). Gut 1988; 29: A1443. 50. Hawkey CJ. Non-steroidal anti-inflammatory drugs and peptic ulcer. Br Med J 1990; 300: 2i8-84. 51. Jick SS, Perera DR, Walker AM, Jick H.Non-steroidal antiinflammatory drugs and hospital admission for perforated peptic ulcer. Lancet 1987; ii: 380-2. 52. Quader K, Logan RFA. Peptic ulcer ( I T ) deaths; how many occur at home or after non-steroidal anti-inflammatory drug (NSAID) prescribing? Gut 1988; 29: A1443. 53. Langman MJ. Epidemiological evidence on the association between peptic ulceration and anti-inflammatory drug use. Gastro Enterol 1989; 96 (Suppl): 640-6. 54. Beard K, Walker :LU, Perera DR, Jick H. Non-steroidal antiintlammatory drug and hospitalisation for gastroesophageal bleeding in the elderly. Arch Intern Med 1987; 147: 1621-3. 55. I'angman MJS, Brooks P, Hawkey CJ, Silverstein F, Yeomans N. Non-steroidal anti-inflammatory drug associated ulcer: epidemiology, causation and treatment. J Gastroenterol Hepatol 1991; 6: 442-9. 56. Laporte JR, Came X, Vidal X, Moreno V, Juan J. LJpper gastrointestinal bleeding in relation to previous use of analgesics and non-steroidal anti-inflammatory drugs. Lancet 1991; 337: 85-9. 57. Fries JF, Miller SR, Spitz PW, Williams CA, Hubert HB, Bloch DA. Toward an epidemiology of gastropathy associated with non-steroidal anti-inflammatorydrug use. Gastroenterology 1989; 96: 647-55. 58. Piper JM, Ray Wh, Daugherty JR, Griffin MK. Corticosteroid use and peptic ulcer disease: role of non-steroidal antiinflammatory drugs. Ann Intern Med 1991; 114: 735-40. 59. lries JF, Williams CA, Bloch DA. The relative toxicity ofnonsteroidal anti-inflammatory drugs. Arthritis Rheum 1991; 34: 1353-60. 60. Henry DA, Johnson A, Dobson A, Iluggan J. Fatal peptic ulcer complications and the use of' non-steroidal anti-inflammatory drugs, aspirin and corticosteroids. Br Med J 1987; 295: 1227-9. 61. Bijlsma JWJ. Treatment of MAID-induced gastrointestinal lesions with cimetidine: an international multicentre collaborative study. Aliment Pharmacol Ther 1982; 2 (Suppl 1): 89-96. 62. Croker JR, Cotton PR, Boyle AC, Kinsella P. Cimetidine for peptic ulcer in patients with arthritis. Ann Rheum Dis 1980; 39: 275-8. 63. lhvies J, Collins AJ, Dixon A StJ. T h e influence of cimetidine on peptic ulcer in patients with arthritis taking anti-inflammatory drugs. Br J Rheumatol 1986; 25: 54-8. 64. Farah D, Sturrock RI), Russell RI. Peptic ulcer in rheumatoid arthritis. Ann Rheum Dis 1988; 47: 478-80. 65. Walan A, Rader JP, Classen M, Lamaers CB, Piper DW. Effect of omeprazole and ranitidine on ulcer healing and relapse rates in patients with benign gastric ulcer. N Engl J Med 1989; 320: 69-75. 66. 1.ancaster-Smith MJ, Jaderberg ME, Jackson DA. Ranitidine in the treatment of NSAID-associated gastric and duodenal ulcers. Gut 1991; 32: 252-5. 67. Isenberg JI, Peterson WL, Elashoff JD er al. Healing of benign gastric ulcer with low-dose antacid or cimetidine. A double-blind, randomized, placebo-controlled trial. N Engl J Med 1983; 308: 1319-24. RROOKS AND YEOiMANS
68. Caldwell JR, Roth SH, Wu WC p r a/. Sucralfate treatment of non-steroidal anti-inflammatory drug-induced gastrointestinal symptoms and mucosal damage. Am J Med 1987; 8 3 (Suppl 3B): 74-82. 69. Malchow-Moller A, Ranlov PJ. Does sucralfate reduce acetylsalicylic acid-induced gastric mucosal bleeding? Scdnd J Gastroenterol 1987; 22: 550-2. 70. Rorh S, Agrawal N, Mahowald M ~f a/. Misoprostol heals gastroduodenal injury in patients with rheumatoid arthritis receiving aspirin. Arch Intern Med 1989; 149: 775-9. 71. Paladini G, Fabiani MG, Tosato F, Maggiolo F. Prophylactic and therapeutic role of rioprostil in NSAIDs induced gastroduodenal lesions. Scand J Gastroenterol 1989; 224 (Suppl 164): 242-6. 72. Sontag SJ, Schnell TG, Mak E el a/. Enprosti1 heals NSAID induced gastric ulcers. Gastroenterology 1990; 98: A1 29. 73. Euler AR, Safdi M, Rao j er QI. A report of three multiclinic trials evaluating arhaprostil in arthritic patients with ASANSAID gastric mucosal damage. Gastroenterology 1990; 98: 1549-57. 74. Shearman DJC, Finlayson NDC. Diseases of the gastrointestinal tract and liver, 2nd ed. Edinburgh: Churchill Livingstune 1989. 75. Ehsanalluh RSB, Page MC, Tildesley G er al. Prevention of gastroduodenal damage induced by non-steroidal anti-inflammatory drugs: controlled trial ofranitidine. Br Med J 1988; 297: 1017-21. 76. Kobinson MG, GrifFin JW Jr, Rowers J, Kogan Fj er al. Effect of ranitidine on gastroduodenal mucosal damage induced by nonsteroidal anti-inflammatory drugs. Dig Dis Sci 1989; 34: 424-8. 77. Berkowitz JM, Rogenes PR, Sharp J?',Warner CW. Ranitidine protects against gastroduodenal mucosal damage associated with chronic aspirin therapy. Arch Intern Med 1987; 147: 2137-9. 78. Geis GS, Stead H, Wallemark CB, Nicholson PA. Prevalence of mucosal lesions in the stomach and duodenum due to chronic use of NSAID in patients with rheumatoid arthritis or osteoarthritis and interim report on prevention of diclofenac associated lesions. J Rheumatol Suppl 1991; 28: 11-4.
NONSTEROIDAL ANTI-1NFI.AILIMATORY DRUGS AND GI
79. Graham IIU, Agrawal NM, Roth SH. Prevention of NSAIDinduced gastric ulcer with rnisoprostol multicentre double blind placebo controlled trial. Lancet 1988; ii: 1277-80. 80. Elliott SL, Yeonians ND, Buchanan RRC er 11. Long term effects of misoprostol on gastropathy induced by non-steroidal antiinflammatory drugs (NSAIIh). Gastroenterology 1990; 8: S59. 81. Agrawal N, Stromatt S, Brown J. Comparative study of misoprostol and sucralfate in rhe prevention of NSAID induced gastric ulcers. Clin Exp Rheuniatol 1990; 8: 59. 82. Parr G, Darekar B, Fletcher A, Rulpitt CJ. Joint pain and quality of life: results of a randomised trial. Br J Clin Pharmacol 1989; 27: 235-42. 83. Armstrong CP, Blower AL. Non-steroidal anti-inflammatory drugs and life threatening complications of peptic ulceration. Gut 1987; 28: 527-32. 84. Editorial. Topical NSAIDs: a gimmick or a godsend. Lancet 1989; 779-80. 85. Gabriel SE. Is misoprostol prophylaxis indicated for NSAID induced adverse gastrointestinal events? An epidemiologic opinion. J Rheumatol 1991; 18: 958-61. 86. Hillman AI,, Bloom BS. Economic effects of prophylactic use of misoprostol to prevent gastric ulcer in patients taking nonsteroidal anti-inflammatory drugs. Arch Intern Med 1989; 149: 2061-5. 87. Knill-Jones R, Drumniond M, Kohli H, Davies I,. Economic evaluation of gastric ulcer prophylaxis in patients with arthritis receiving non-steroidal anti-inflammatory drugs. Postgrad Med J 1990; 66: 639-46. 88. Edelson JT, Tosreson ANA, Sax P. Cost effectiveness of misoprostol for prophylaxis against non-aeroidal anti-inflammatory druginduced gastrointestinal bleeding. JAMA 1990; 264: 41-7. 89. Barrison I. Prophylaxis against non-steroidal induced upper gastrointestinal side effects. Ann Rheum Dis 1991; 5 0 207-9.
Aust NZ J Med 1992; 22
691
