16 (2006) 83–97

Drug Injury in the Upper Gastrointestinal Tract:

Nonsteroidal Anti-Inf lammatory Drugs

Mae F. Go, MD

Veterans Administration Salt Lake City Health Care System, 500 Foothill Boulevard,

111G, Salt Lake City, UT 84106, USA

Nonsteroidal anti-inflammatory drugs (NSAIDs), including aspirin, are

probably the most commonly used class of drugs. The frequency of NSAID-

related upper gastrointestinal complications is low—estimated at approximately

1% to 2% of all NSAID users, with an overall risk for gastrointestinal com-

plications of 1% to 4% per year. However, the huge number of individuals who

consume these drugs (an estimated 60 million Americans in the year 2000) results

in a significant burden of illness [1]. Adverse effects of NSAIDs and aspirin have

been recognized for years and range from dyspepsia without endoscopic findings

to severe complications of ulcer, such as bleeding and perforation.

The large number of NSAIDs and their ready availability over the counter

(OTC) makes it difficult to know the exact quantity of these drugs being used.

One hundred million prescriptions for NSAIDs were written in the year 2000,

with an estimated $1.8 billion spent on OTC aspirin and nonaspirin NSAIDs.

These drugs are most commonly used for pain syndromes and rheumatic

disorders. More than 40% of the population older than 65 years are prescribed

NSAIDs, and 1% to 8% are hospitalized for an NSAID-related gastrointestinal

complication [2]. Adverse events result in more than 100,000 hospitalizations

each year. Approximately 17,000 deaths are attributed to NSAID and aspirin use.

Recent studies support the prophylactic use of these drugs (particularly aspirin)

for many clinical diseases, such as cardiovascular disease, stroke, and possibly

breast cancer [3–5]. Many trials suggest that aspirin and NSAID use may

decrease colon polyp recurrence [6]. These reports have resulted in enormous

numbers of patients’ using these drugs. The recent withdrawal of cyclo-

oxygenase 2 (COX-2) selective inhibitors from global markets exemplifies the

Gastrointest Endoscopy Clin N Am

1052-5157/06/$ – see front matter. Published by Elsevier Inc.

doi:10.1016/j.giec.2006.01.001 giendo.theclinics.com

E-mail address: mae.go@med.va.gov

go84

renewed interest in and importance of awareness of NSAID-related complica-

tions, as well as understanding of how to manage and prevent them. The US Food

and Drug Administration (FDA) has requested that manufacturers of prescription

and OTC NSAIDs revise their product package inserts to include boxed warnings

regarding potential cardiovascular and gastrointestinal risks [7].

Pathogenesis of gastrointestinal mucosal injury

The increased risk for NSAID gastrointestinal toxicity has been attributed to

decreased prostaglandin synthesis and to the increased number of comorbidities

that predispose to (increase risk for) NSAID-related complications, especially at

an advanced age. NSAID-induced mucosal injury most likely occurs through the

failure of multiple gastroprotective mechanisms to prevent a reduction in mucosal

blood flow at the submucosal and mesenteric levels (for examples, see [8,9]).

This mucosal injury occurs by two mechanisms: topical and systemic. Topical

injury results from increased cellular permeability with diffusion of the drugs into

mucosal cells in the gastric acidic environment. High intracellular concentrations

result from ion-trapping in the neutral intracellular pH. NSAIDs also reduce

cellular phospholipids and gastric mucus surface hydrophobicity, increasing

susceptibility to noxious agents [10].

Systemic mechanisms of NSAID injury involve decreased synthesis of

protective gastric mucosal prostaglandins. NSAIDs inhibit cyclo-oxygenase 1

(COX-1) and COX-2 activity, preventing generation of prostaglandins from

arachidonic acid. Inhibition of COX-2 provides the beneficial effects of the

NSAIDs by reducing inflammation. COX-1 inhibition mediates decreased gastric

mucin and bicarbonate synthesis, reduced mucosal blood flow, and decreased

cellular regeneration. NSAIDs with greater inhibition of COX-1 have been

associated with greater risk for serious gastrointestinal complications [11].

Vascular injury may also result from an increase in adhesion molecules and

neutrophil adherence with stasis and ischemia. NSAIDs can also increase

pepsinogen secretion and affect acid secretion.

Risks related to nonsteroidal anti-inflammatory drugs and aspirin

The reported magnitude of risk contributed by NSAIDs and aspirin differs in

studies depending on the population studied, the study design, and the particular

NSAID or NSAIDs used by the study population. The estimated risk ratio (RR)

for perforations, ulcers, and bleeds associated with NSAID use varied from odds

ratio (OR) 5.36 (95% confidence interval [CI], 1.79–16.1) in randomized,

controlled trials to RR 2.7 (95% CI, 2.1–3.5) in cohort studies to OR 3.0

(95% CI, 2.5–3.7) in case-control studies [12].

The clinical presentation of NSAID gastrointestinal toxicity is variable and

includes gastric or duodenal erosions, gastric and duodenal ulcers, ulcer bleeding,

nonsteroidal anti-inflammatory drug injury 85

gastric outlet obstruction, and perforation. Gastric ulcers are reported in 15% to

20% of NSAID users, whereas duodenal ulcers occur in 5% to 8%. Diffuse

gastric hemorrhage has also been attributed to NSAID use. An estimated 15% to

25% of NSAID users will seek medical help for dyspepsia.

Poor correlation exists between symptoms and clinical complications. Absence

of symptoms may be related to masking of symptoms by the analgesic properties

of the NSAIDs. The first sign of an ulcer was a life-threatening complication in

58% of patients using NSAIDs in one study [13]. An endoscopic study of asymp-

tomatic patients on ‘‘minidose’’ aspirin revealed ulcer or erosions in 48% [14].

Dyspepsia tends to be a poor marker for structural disease (endoscopic

mucosal damage). Risk for dyspepsia is increased by 36% in NSAID users [15].

A population-based study in Olmstead County, Minnesota reported that aspirin

and nonaspirin NSAID use was associated with an almost twofold increased risk

of upper gastrointestinal symptoms in the population 65 years and older [16].

Adjusted dyspepsia prevalence was 15.0 per 100 (95% CI, 12.9–17.0) and was

similar across age groups and in men and women.

Multiple factors increase risk for NSAID-related complications (Table 1).

Greater age consistently appears as an independent factor for increased risk. A

meta-analysis of cohort studies found a 0.1% 1-year prevalence of serious

gastrointestinal complications among all NSAID users [17]. The 1-year

prevalence among NSAID users who were 65 years and older was 0.32%,

compared with 0.039% in those NSAID users who were younger than 65 years.

Increasing age increases relative risk. In patients 60 to 75 years of age, relative

risk was 3.5 (95% CI, 2.5–4.1), but it increased to RR 8.9 (95% CI, 4.0–10.2) in

those older than 75 years [9].

Duration and dose of the NSAID may affect the risk for upper gastrointestinal

complication outcomes. In a case-control study, upper gastrointestinal bleeding

risk was similar for aspirin at any dose and NSAIDs when the medication was

taken for less than 1 month (OR 8.9 [95% CI, 6.6–14.5] for aspirin and OR 7.6

Table 1

Factors increasing risk for peptic ulcer disease related to nonsteroidal anti-inflammatory drugs

complications

Factor Risk

Age N 60–75 y RR 3.5 [8]

Age N 75 y RR 8.9 [8]

NSAID plus aspirin SIR 5.6 [20]

Prior history of one complicated PUD RR 4.8 [16]

High-dose NSAIDs OR 9.4 [17]

Concomitant warfarin RR 6.4 [8]

Concomitant steroids RR 2.2 to 4.4 [8,18]

Aspirin (dose-related) RR 2.5 to 3.9 [22]

SSRI + NSAID RR 2.9 to N12 [33–35]

SSRI + low-dose aspirin RR 5.2 [33,35]

Abbreviations: PUD, peptic ulcer disease; SIR, standardized incidence ratio; SSRI, selective serotonin

reuptake inhibitor.

go86

[95% CI, 5.8–9.2] for NSAIDs) [18]. Risk decreased when the drugs were taken

for longer than 1 month, to OR 2.8 (95% CI, 1.6–4.0) for aspirin and OR 2.4

(95% CI, 1.8–9.2) for NSAIDs. ‘‘Standard-dose’’ NSAID intake increased risk

compared with that of nonusers to OR 6.9 (95% CI, 6.0–9.3), whereas taking

higher doses of NSAIDs increased it to OR 9.4 (95% CI, 6.3–14.2). Notably,

concurrent use of antisecretory medications decreased risk (OR 0.6 [95% CI,

0.4–0.9]). A prior NSAID-related event significantly increased RR to 4.76

(CI, 4.05–5.59) for another NSAID complication [17]. When patients 60 years

and older using NSAIDs were compared with nonusers, those using NSAIDs had

increased risk of 5.52 (95% CI, 4.63–6.6).

A prior complicated ulcer conferred a relative risk of 4.8 (95% CI, 4.0–5.6) in

a meta-analysis of case-control and cohort studies [17]. Relative risks of con-

comitant steroids vary from RR 2.2 (95% CI, 1.8–14.6) [9] to RR 4.4 (95% CI,

2.0–9.7) [19]. Concurrent NSAID and anticoagulant use had RR 6.4 (95% CI,

2.8–14.6). A nested case-control study analyzed multiple health care data-

bases in Ontario, Canada for association between use of warfarin and NSAIDs

(including COX-2 inhibitors) and risk for gastrointestinal bleeding in patients

older than 66 years [20]. These authors found similar risks of gastrointestinal

bleeding for NSAIDs and COX-2 inhibitors: OR 1.9 (95% CI, 1.4–3.7) and OR

1.7 (95% CI, 1.2–3.6), respectively. They found no differences in risk based on

gender or symptom.

Aspirin has been considered a benign medication, but many studies now

confirm significant gastrointestinal risks with this drug. The standardized

incidence ratio (SIR) for individuals on low-dose aspirin alone (100–150 mg

daily) in a Danish cohort study was 2.6 (95% CI, 2.2–2.9), whereas those taking

both low-dose aspirin and an NSAID had a SIR of 5.6 (95% CI, 4.4–7.0) [21].

Buffering and enteric-coating aspirin do not decrease risks for upper gastro-

intestinal bleeding compared with plain aspirin. Kelly and colleagues [22]

calculated that the relative risks for upper gastrointestinal bleeding for plain,

buffered, and enteric-coated aspirin at less than 325 mg daily were 2.6, 3.1, and

2.7, respectively. Weil and colleagues [23] found a dose-response relationship

between aspirin and upper gastrointestinal bleeding; 75 mg had an OR of 2.3

(CI, 1.2–4.4), 150 mg had an OR of 3.2 (CI, 1.7–6.5), and 300 mg had an OR of

3.9 (CI, 2.5–6.3) in patients who were older than 60 years.

NSAID users have a two- to threefold increased risk for gastrointestinal

bleeding and a fivefold increased risk for gastrointestinal-related mortality

[11,17]. A recent study of the Spanish population from Lanas and colleagues [24]

suggests that risks may have decreased. A population-based observational study

in Spain quantified NSAID gastrointestinal toxicity and mortality in hospitals

providing care for 80% of the population using two data sets representative of

common clinical practice. The authors detected 15.3 deaths due to NSAID-

related gastrointestinal complications per 100,000 NSAID users, an estimate that

is much lower than those in the United States and the United Kingdom. The

authors had previously suggested the possibility of decreased NSAID compli-

cations when patients are taking concurrent antisecretory drugs [18]. They

nonsteroidal anti-inflammatory drug injury 87

indicate in their recent study that as many as one third of all NSAID-related

deaths in their population are attributed to low-dose aspirin.

Is gender a risk factor for complications related to nonsteroidal

anti-inflammatory drugs?

Gender and response to pharmaceutic agents is a controversial issue (for

example, see [25]). In contrast to studies in men’s health, the Women’s Health

Study reported that aspirin may not decrease risk for cardiovascular disease in

women (relative risk 1.02; 95% CI, 0.84–1.25) and has little effect on all-cause

mortality (relative risk 0.95; 95% CI, 0.85–1.06) [4,26].

Straus and Ofman [11] reported that the pooled relative risk for NSAID

complications was greater in women (RR ~5) than in men (RR ~3) when NSAID

users were compared with non–NSAID users. They suggested that comorbidities

or patterns of drug use might account for this gender difference. Henry and

colleagues [27] suggested that women older than 65 years had greater risk, but a

recent meta-analysis showed no difference in NSAID-related complications based

on gender [17]. Irrespective of whether gender influences risk for NSAID-related

gastrointestinal complications, consideration must be given to the proportions of

men and women in the United States population to gauge the overall burden of

illness related to complications from NSAIDs.

The United States Census 2000 reported that, of the 282.4 million people in

the United States population, 50.9% were women [28]. These data suggest a

nearly equal male/female ratio in the general population. It is enlightening to

examine the male/female ratio (the greater the number, the more men compared

with women). The Census 2000 reported that, up through the 25-to-34-year-old

group, the male/female ratio is greater than 100, reflecting the larger number of

boys born every year and the greater number of men in early childhood and

adulthood. However, the male/female ratio decreases rapidly in older age-groups,

with women outnumbering men. In the age group of 55 to 64 years, the male/

female ratio is 92, and it drops to 41 in the 85-and-older age group. With

advancing age, women significantly outnumber men, particularly when comorbid

diseases, such as rheumatic diseases with joint symptoms, become more

prevalent. Seventy percent of people older than 70 years take NSAIDs. Forty

million Americans had some form of arthritis in 1995. Twenty-one million

American adults have osteoarthritis, and 60% of arthritis patients are women [29].

Gender-related issues become particularly important when considering the older

patient, who is more likely to be female with comorbidities.

Age and increased risk for nonsteroidal anti-inflammatory drug–related

complications

Until the discovery of Helicobacter pylori, it was believed that gastric acid

secretion declined with age, but recent studies indicate that aging alone (without

H pylori infection) does not predispose to diminished acid secretion. Detailed

go88

reviews of NSAID gastrointestinal toxicity have been published [8,9,29]. NSAID

toxicity is significantly greater in patients older than 60 years, with an estimated

relative risk of 5.5 (95% CI, 4.6–6.6) compared with 1.7 in younger subjects [9].

Relative risk was 3.5 in patients 60 to 75 years of age, compared with 8.9 in

patients older than 75 years. Although the prevalence of peptic ulcer has de-

clined, the prevalence of ulcer complications in the older population continues to

rise [30].

Animal and human studies have helped to define potential mechanisms of

age-related susceptibility to NSAID injury. Lipscomb and colleagues [31] per-

formed a retrospective analysis of 62 healthy volunteers to examine factors

related to gastric adaptation after a 28-day course of either etodolac 300 mg twice

a day, naproxen 500 mg twice a day, or diclofenac 50 mg twice a day. Modified

Lanza score at endoscopy on days 1, 7, and 28 was used to assess gastric mucosal

damage. During the 4-week study, 42 subjects (71%) adapted completely,

13 (22%) had incomplete adaptation, and four (7%) had no evidence of adap-

tation. An age-related response was found, with a trend toward more adap-

tation in younger subjects. Patients who did not achieve adaptation had a

mean age of 42.0 F15.7 years; those with a partial response had a mean age of

32.5 F10.3 years, and those with complete adaptation had a mean age of

26.8 F9.8 years.

Tarnawski and colleagues [32] found 58% decreased blood flow, fourfold

reduced prostaglandin E2 generation, and 5.2-fold increased connective tissue

around microvessels in older rats compared with young rats at baseline. Challenge

with indomethacin in the older rats caused earlier and greater reduction in mucosal

blood flow (at 2 hours, down to 25% of normal value in young rats), increased

tissue hypoxia, and a further reduction in prostaglandin E2 generation.

The United States Census 2000 reported that the percentage of people 65 years

and older remained the same, at 12% of the total population, but there was a 12%

increase in the absolute number, from 31.2 million in 1990 to 35 million in 2000.

Furthermore, the number of people who were 85 years or older increased by 38%

in this same period, from 3.1 million to 4.2 million; the 75- to 84-year-old group

increased by 23%.

Over the counter versus prescription nonsteroidal anti-inflammatory drugs:

is there a difference in risk?

Lewis and colleagues [33] conducted a case-control study of upper gastro-

intestinal toxicity related to OTC NSAIDs. Three hundred and thirty-nine case

subjects hospitalized for upper gastrointestinal bleeding, perforation, or non-

malignant gastric outlet obstruction were compared with 1889 controls. A 1.8

adjusted OR in cases versus controls was seen with only 4- to 7-day use of an

OTC NSAID, which is similar to rates seen with prescription NSAIDs. Risks of

high-dose OTC NSAIDs were similar to those of prescription NSAIDs (adjusted

OR 5.21). As many as a third of individuals taking prescription NSAIDs were

also taking OTC NSAIDs.

nonsteroidal anti-inflammatory drug injury 89

Serotonin reuptake inhibitors, nonsteroidal anti-inflammatory drugs, and

gastrointestinal bleeding

Retrospective studies have reported an increased risk for gastrointestinal

bleeding in patients who are taking selective serotonin reuptake inhibitors

(SSRIs). Concurrent exposure to NSAIDs and SSRIs can increase risk of

gastrointestinal bleeding by 12- to 15-fold [34,35]. de Abajo and colleagues [35]

reported a relative risk of 3.7 for upper gastrointestinal bleeding in individuals

using NSAIDs alone, which increased to 15.6 with concurrent use of SSRIs.

They found no increased risk for upper gastrointestinal bleeding when non-SSRI

antidepressants were used, and no difference in risk based on gender or age.

Dalton and colleagues [34] examined the association of antidepressant use and

hospitalization for upper gastrointestinal bleeding in the Danish population. More

than two thirds of the individuals using antidepressants during the study period

were women, and 51% were younger than 59 years. When SSRIs were used alone,

the RR was 3.6 (95% CI, 2.7–4.7), but it increased to 12.2 (95% CI, 7.1–19.5)

when they were combined with NSAID use. The addition of low-dose aspirin to

SSRI use increased the RR to 5.2 (95% CI, 3.2–8.0). Even the use of non-SSRI

antidepressants with low-dose aspirin or an NSAID increased RR to 4.6 and 8.2,

respectively. Another population-based case-control analysis of computerized

primary care data (the Health Improvement Network database) calculated risk for

gastrointestinal bleeding in patients taking concurrent SSRI and NSAID

medications [36]. The authors examined 11,261 patients who had initial

gastrointestinal bleed and 53,156 matched controls. They calculated an OR of

2.09 (95% CI, 1.91–2.28) for any antidepressant exposure and first gastrointestinal

bleed, an OR of 2.38 (95% CI, 2.08 – 2.72) for SSRI and first gastrointestinal

bleed, an OR of 1.87 (95% CI, 1.67 – 2.09) for any tricyclic antidepressant, and an

OR of 2.71 (95% CI, 2.50 – 2.94) for any nonselective NSAID. The OR of 2.93

(95% CI, 2.39 – 3.34) was greater with concurrent exposure to SSRI medication

and an NSAID compared with single-drug exposure. Any antidepressant

combined with NSAID use had an OR of 3.8 (95% CI, 2.91–4.96). Interestingly,

the authors found no increased risk for gastrointestinal bleeding in subjects older

than 80 years. Cotherapy with gastroprotective drugs did not alter risks with

single-agent use or concurrent use of antidepressants and NSAIDs. Although the

risk was not as high as reported in earlier studies, use of antidepressants does

appear to confer increased risk for gastrointestinal bleeding.

Takeuchi and colleagues [37] examined the effects of combined SSRI and

indomethacin use in a rat model. SSRIs alone did not directly cause mucosal

lesions. Indomethacin caused gastric corpus lesions in the fasted state, but the

SSRI paroxetine suppressed these indomethacin-induced lesions. By contrast,

antral ulcers were seen in animals fed after a fasting period; antral ulcers were

aggravated with the addition of an SSRI, and many lesions progressed from

nonbleeding to bleeding lesions.

Platelet aggregation is augmented by serotonin release from the platelet. If the

SSRIs prevent reuptake and storage of serotonin, increased gastrointestinal

go90

bleeding may result. These studies suggest that any antidepressant may increase

risk for gastrointestinal bleeding, so factors other than serotonin must be crucial

in inciting gastrointestinal bleeding. A limitation of all these studies is that

detailed clinical information was not available to confirm gastrointestinal

bleeding and its source. Nonetheless, they indicate that the list of drugs that

can cause gastrointestinal tract bleeding continues to increase.

Management strategies for the patient who needs nonsteroidal

anti-inflammatory drugs

The optimal therapeutic strategy should be guided by individual indications

and risk factors (Box 1). The first requirement is to determine indications for

Box 1. Strategies for managing ulcers and gastropathy induced bynonsteroidal anti-inflammatory drugs

A. Treatment

1. If possible, discontinue the NSAID/aspirin.2. Heal the ulcer with acid suppression (PPIs are superior

to H2RAs).3. Prevent ulcer recurrence.

a. Eradicate H pylori infection if present.b. Consider prophylactic cotherapy with PPI if NSAID

is continued.c. Determine whether non-NSAID analgesics may

be substituted.4. Remember factors conferring high risk.

B. Prevention of NSAID- or aspirin-associated complications

1. Assess indication or indications.2. Consider non-NSAID agents.3. Ask about high-risk factors.4. Select lowest possible NSAID dose.5. Avoid multiple NSAIDs.6. Consider PPI cotherapy for significant risks.7. Inform patient of potential risks.*

* Patient information is available at Web sites for the US FDA,American Gastroenterological Association, and American College ofGastroenterology. Boxed warnings should accompany prescriptionand OTC NSAIDs.

nonsteroidal anti-inflammatory drug injury 91

NSAID use. If NSAIDs are needed, for what duration and dose? What are the

patient’s risks for a complication? The need for primary or secondary prophylaxis

should be considered. Medical therapies include the prostaglandin analogue

misoprostol and acid-suppressing agents, including histamine-2 receptor antago-

nists (H2RAs) and proton pump inhibitors (PPIs). In the patient who has had an

NSAID or aspirin clinical event, the first step should be to discontinue the

NSAID or aspirin. When the causative agent is discontinued, high healing rates

may be achieved with antisecretory therapy.

If the NSAID or aspirin cannot be discontinued, a reduced dose or a ‘‘less

toxic’’ NSAID may be considered [8,9,11]. Cotherapy with a gastroprotective

agent such as the PPIs should be considered. Misoprostol at full dose also has

high healing and preventive rates, but the high frequency of diarrhea leading to

noncompliance makes this approach less desirable.

Peptic ulcer healing rates with H2RAs have been reported as high as 80% to

95% at 8 weeks of treatment. Lancaster-Smith and colleagues [38] observed 95%

and 100% 8-week healing rates for gastric ulcers and duodenal ulcers,

respectively, in patients who discontinued NSAIDs. Healing rates were much

lower when patients continued NSAIDs (63% and 84% for gastric ulcer and

duodenal ulcer healing, respectively). The efficacy of ranitidine at a standard dose

(150 mg twice a day) compared with placebo in preventing upper gastrointestinal

mucosal damage by NSAIDs was evaluated in an 8-week endoscopic study of

144 patients on chronic NSAIDs [39]. No difference was found in gastric ulcer

development between the arms (10% to 12%). However, 8% of patients in the

placebo arm developed a duodenal ulcer compared with none in the ranitidine

arm, suggesting that ranitidine may prevent duodenal ulcers, although it had no

advantage in preventing gastric ulcers compared with placebo.

The Misoprostal Ulcer Complications Outcomes and Safety Assessment

(MUCOSA) trial compared the efficacy of misoprostol 200 mg four times daily

with placebo for prevention of gastrointestinal complications in patients who

had rheumatoid arthritis and were continuing NSAID use [40]. At the 6-month

follow-up, 0.74% of patients on placebo had serious complications, compared

with 0.36% of those receiving misoprostol (P = .049; RR 0.46). Misoprostol is

effective for prevention of gastrointestinal complications at full doses, but the

frequent development of diarrhea often leads to discontinuation of the drug

(see following discussion).

Raskin and colleagues [41] evaluated ulcer prevention in patients using

chronic NSAIDs who were randomized to misoprostol 200 mg four times daily

or ranitidine 150 mg twice daily in an 8-week study. Duodenal ulcer develop-

ment was similar in both groups (1%), but patients randomized to misoprostol

had significantly fewer gastric ulcers compared with those taking ranitidine

(0.56% versus 5.6%; P b.01). Side effects resulting in study withdrawal were

twice as high in the misoprostol arm as in the ranitidine arm (13% and

6.7%, respectively).

Ekstrom and colleagues [42] performed one of the first studies to compare PPI

with placebo for prevention of NSAID-related ulcers. During the 3-month study,

go92

4.7% of omeprazole-treated patients developed an ulcer (2% gastric and 2%

duodenal), compared with 16.7% (7% gastric and 10% duodenal) of those

randomized to placebo. Dyspepsia developed in 15% of those in the omeprazole

group, compared with 36% in the placebo group. At the end of the 3-month study,

74% of patients in the omeprazole arm were in remission, compared with only

48% in the placebo arm. Similar results were seen by Cullen and colleagues [43]

in their 6-month study comparing omeprazole 20 mg with placebo once daily for

primary prophylaxis of NSAID ulcer development in patients continuing NSAID

therapy. In the placebo arm, 16.5% of patients developed ulcers, compared with

3.6% of those in the omeprazole arm.

Hawkey and colleagues [44] conducted a healing and maintenance study in

935 patients on chronic NSAIDs. Patients were randomized to omeprazole 20 mg

or 40 mg daily or misoprostol 200 mg four times a day; ulcer healing at 8 weeks

was similar in all groups (71% to 75%). Seven hundred and thirty-two patients

with healed ulcers were randomized to 6-month maintenance treatment with

omeprazole 20 mg once daily, misoprostol 200 mg twice daily, or placebo to

assess ulcer prevention while the patients continued NSAIDs. More patients

maintained on omeprazole were in remission (61%) compared with those on

misoprostol (48%; P = .001) or placebo (27%; P b.001). It should be noted that

misoprostol was used at lower than approved doses.

Yeomans and colleagues [45] compared omeprazole 20 mg/d with ranitidine

150 mg twice daily for NSAID-related ulcer healing, followed by a maintenance

phase with both drugs to assess 6-month remission rates, in 541 patients

continuing NSAIDs. The 8-week gastric ulcer healing rate with omeprazole was

84%, versus 64% with ranitidine (P b.001). Duodenal ulcer healing was 92%

with omeprazole versus 81% with ranitidine (P = .03). At the 6-month follow-up,

72% of patients randomized to omeprazole remained in remission, compared with

59% of those in the ranitidine arm.

PPI efficacy for NSAID-related upper gastrointestinal symptom alleviation has

recently been examined. An international multicenter study compared 6-month

maintenance treatment with omeprazole 20 mg daily, misoprostol 200 mg twice

daily, and placebo for secondary prevention of NSAID-related dyspepsia in

arthritis patients [46]. Quality of life scores improved with both omeprazole and

misoprostol. However, when patients were treated with omeprazole, gastro-

intestinal symptoms were alleviated or remained at the alleviated level after ulcer

healing, whereas abdominal pain and indigestion symptoms significantly

worsened in patients in the misoprostol treatment arm. These patients also

suffered significant diarrhea, despite the lower dose.

To confirm PPI efficacy for NSAID-related symptom alleviation, esomeprazole

was evaluated in symptomatic patients who were confirmed to have no ulcer

disease [47]. Twomultinational trials examined adults who had a chronic condition

requiring NSAIDs for longer than 7 months; NSAIDs included nonselective

NSAIDs, selective COX-2 inhibitors, high-dose aspirin (N 325 mg/d), or a

combination. The investigators randomized 1156 patients who had moderate

dyspepsia with negative baseline upper endoscopy and negative H pylori tests to

nonsteroidal anti-inflammatory drug injury 93

placebo or esomeprazole 20 mg/d or 40 mg/d for 4 weeks. Esomeprazole at either

dose was significantly more effective than placebo in alleviating upper gastro-

intestinal symptoms compared with baseline. Symptom improvement on esome-

prazole 20 mg was 2.3 mean and 2.17 in each study, respectively, compared with

placebo, with symptom improvement scores of 1.64 and 1.56 in each study.

Prevention of NSAID-related lesions and symptoms using pantoprazole 20 mg

once daily was compared with use of misoprostol 200 mg twice daily in

515 high-risk rheumatic patients [48]. Remission rates at 6 months by intention-

to-treat analysis (ulcer, N10 erosions, or reflux esophagitis) were 89% and 70%

for pantoprazole and misoprostol, respectively (95% CI, 11%–27%; P = .005).

Patients in ‘‘symptomatic’’ remission at 6 months included 99% and 92% for

pantoprazole and misoprostol, respectively (P = .002). The total number of

gastroduodenal lesions may have been underestimated in this study because

patients with 10 or more erosions at the 3-month endoscopy were withdrawn.

Recent clinical trials include patients who have negative H pylori tests to

eliminate possible confounding effects of the infection in peptic ulcer healing

and remission. A 12-week double-blind, endoscopic study was conducted

in 537 patients continuing chronic NSAIDs with a documented gastric ulcer

and negative tests for H pylori infection [49]. Once ulcers healed, patients

were randomized to receive placebo, misoprostol 200 mg four times a day, or

lansoprazole 15 mg or 30 mg once daily for 12 weeks. By week 12, the pro-

portions of intent-to-treat patients remaining in remission (without gastric or

duodenal ulcer disease) were 47% in the placebo group (95% CI, 41.1% – 61%),

88% in the misoprostol group (95% CI, 87.2% – 97.9%), 79% in the lansoprazole

15-mg group (95% CI, 72.5% – 87.3%), and 83% in the lansoprazole 30-mg

group (95% CI, 75%– 89.6%). Patients receiving lansoprazole or misoprostol

had significantly lower ulcer recurrence than those receiving placebo, but 22%

of misoprostol patients had diarrhea, compared with 7% of those receiving

lansoprazole 30 mg and 3% of those receiving lansoprazole 15 mg.

The PPI esomeprazole has also been effective in reducing gastric and duodenal

ulcers in at-risk patients taking continuous NSAIDs or COX-2 selective

inhibitors. The number needed to treat to prevent one event was similar for

esomeprazole 20 mg and 40 mg given once daily: nine and eight, respectively

[50]. This finding translated into one ulcer prevented per 6 months in every eight

or nine NSAID-using patients who were treated with once-daily esomeprazole.

Scheiman and colleagues [51] studied 1429 patients taking continuous NSAIDs

or COX-2 selective inhibitors for a chronic condition (mostly arthritis). All

patients had increased risk based on age 60 years or greater or history of peptic

ulcer disease within 5 years. Esomeprazole 20 mg or 40 mg given once daily led

to significantly higher remission rates than did placebo (95% versus 83%,

respectively) during the 6-month study period.

PPIs may provide protection against NSAID-induced damage by factors other

than acid suppression. Sarosiek and colleagues [52] examined the impacts of 7-day

NSAID intake plus placebo versus 7-day NSAID intake with rabeprazole

cotherapy on the gastric mucus-buffer layer in a randomized, cross-over study of

go94

21 healthy, H pylori–negative volunteers. Basal and pentagastrin-stimulated

gastric volume, gastric mucus output, and concentration were significantly de-

creased with naproxen exposure. These deleterious conditions were significantly

alleviated (restored) with PPI cotherapy. Twenty of the 21 subjects with NSAID

exposure had decreased gastric mucin output in basal or stimulated conditions,

whereas 13 had decreased mucin secretion in both basal and stimulated conditions.

The authors speculated that these different phenotypic responses may help define

patients who are at greater risk for NSAID-induced injury.

NSAID-related lesions demonstrate more rapid healing with PPIs than with

H2RAs. Two PPIs have received US FDA approval for NSAID-related ulcer

prophylaxis: lansoprazole for secondary prophylaxis and esomeprazole for

primary prophylaxis. Considerable economic resources are consumed by the

complications of and preventive therapies for NSAID and aspirin use. Spiegel

and colleagues [53] examined cost-effective strategies for patients with various

risks and indications for these medications. Their study concluded that generic

nonselective NSAIDs alone were cost-effective for the patient with low risk.

Cotherapy with a PPI with nonselective NSAID was the preferred strategy for

patients at high risk for cardiovascular or gastrointestinal complications.

Substitution of aspirin with clopidogrel to decrease recurrent ulcer bleeding:

is it effective?

Aspirin at any dose increases risk for ulcer and bleeding, particularly in the

older patient who has already suffered an ulcer. Chan and colleagues [54]

examined recurrent ulcer bleeding in patients who initially presented with ulcer

bleeding related to aspirin. After ulcer healing, 320 patients were randomized to

receive clopidogrel 75 mg daily with esomeprazole, placebo, or aspirin 80 mg

plus esomeprazole 20 mg daily for 12 months. Thirteen patients on clopidogrel

developed a recurrent ulcer bleed, compared with one randomized to aspirin plus

esomeprazole, with a cumulative 12-month incidence of 8.6% (95% CI, 4.1%–

13.1%) for clopidogrel versus 0.7% (95% CI, 0%–2%) for those receiving aspirin

plus esomeprazole. How clopidogrel causes ulcer development is unknown, but it

has been speculated that it may impair healing or exacerbate an existing lesion.

This study underlines the importance of assessing gastrointestinal tract risks

related to antiplatelet therapy.

Summary

It is well established that NSAIDs and aspirin confer significant morbidity and

mortality. The widespread use of these drugs has increased the absolute numbers

of cases of NSAID- or aspirin-related upper gastrointestinal complications. The

older population continues to increase, so an increase in the number of patients

taking NSAIDs and aspirin may be expected. The United States Census 2000

nonsteroidal anti-inflammatory drug injury 95

showed a 38% increase in adults in the older than 85 group. The number of

women in the older groups increased likewise; there were 20.6 million women

aged 65 years or older compared with 14.4 million men. Aspirin is not a benign

medication. The addition of patients taking aspirin for disease prophylaxis has

increased the population at risk for gastrointestinal complications. The list of

drugs with significant gastrointestinal toxicity continues to grow. Emerging data

indicate that antidepressants, such as SSRIs and tricyclic antidepressants, also

increase risk for gastrointestinal bleeding. Multiple factors have been identified

that increase risk for NSAID- and aspirin-related upper gastrointestinal

complications. The highest risks are related to age (N 60 years) and prior

complicated peptic ulcer; additional risk factors include use of multiple NSAIDs

and high-dose NSAIDS. The benefits and risks must be carefully considered

when one prescribes NSAIDs and aspirin. Recent studies have demonstrated

enhanced healing and prevention of NSAID- and aspirin-related gastrointestinal

lesions with proton pump inhibitors.

References

[1] Laine L. The role of proton pump inhibitors in NSAID-associated gastropathy and upper

gastrointestinal symptoms. Rev Gastroenterol Disord 2003;3(Suppl 4):S30–9.

[2] Griffin M. Epidemiology of nonsteroidal anti-inflammatory drug–associated gastrointestinal

injury. Am J Med 1998;104(3A):23S–9S.

[3] Hayden M, Pignone M, Phillips C, et al. Aspirin for the primary prevention of cardiovascular

events: a summary of the evidence for the US Preventive Services Task Force. Ann Intern Med

2002;136:161–72.

[4] Ridker PM, Cook NR, Lee IM, et al. A randomized trial of low-dose aspirin in the primary

prevention of cardiovascular disease in women. N Engl J Med 2005;352:1293–304.

[5] Zhang Y, Coogan PF, Palmer JR, et al. Use of nonsteroidal anti-inflammatory drugs and risk

of breast cancer: the Case-Control Surveillance Study revisited. Am J Epidemiol 2005;162:

165–70.

[6] Sandler RS, Halabi S, Baron JA, et al. A randomized trial of aspirin to prevent colorectal

adenomas in patients with previous colorectal cancer. N Engl J Med 2003;348:883–90.

[7] US Food and Drug Administration. Public health advisory. Available at: http://www.fda.gov/

cder/drug/advisory/cox2.htm. Accessed June 15, 2005.

[8] Cryer B. Mucosal defense and repair: role of prostaglandins in the stomach and duodenum.

Gastrointest Endosc Clin N Am 2001;30:877–94.

[9] Lazzaroni M, Bianchi Porro G. Gastrointestinal side-effects of traditional non-steroidal anti-

inflammatory drugs and new formulations. Aliment Pharmacol Ther 2004;20(Suppl 2):48–58.

[10] Giraud MN, Motta C, Romero JJ, et al. Interaction of indomethacin and naproxen with gastric

surface-active phospholipids: a possible mechanism for the gastric toxicity of nonsteroidal anti-

inflammatory drugs (NSAIDs). Biochem Pharmacol 1999;57:247–54.

[11] Straus WL, Ofman JJ. Gastrointestinal toxicity associated with nonsteroidal anti-inflammatory

drugs. Gastrointest Endosc Clin N Am 2001;30:895–920.

[12] Ofman JJ, Maclean CH, Straus SL, et al. A metaanalysis of severe upper gastrointestinal

complications of nonsteroidal anti-inflammatory drugs. J Rheumatol 2002;29:804–12.

[13] Armstrong CP, Blower AL. Non-steroidal anti-inflammatory drugs and life threatening com-

plications of peptic ulceration. Gut 1987;28:527–32.

[14] Niv Y, Battler A, Abuksis G, et al. Endoscopy in asymptomatic minidose consumers. Dig Dis Sci

2005;50:78–80.

go96

[15] Straus WL, Ofman JJ, MacLean C, et al. NSAIDs cause dyspepsia? A metaanalysis evaluating

alternative dyspepsia definitions. Am J Gastroenterol 2002;97:1951–8.

[16] Talley NJ, Evans JM, Fleming KC, et al. Nonsteroidal anti-inflammatory drugs and dyspepsia in

the elderly. Dig Dis Sci 1995;40:1345–50.

[17] Gabriel S, Jaakkimainen L, Bombardier C. Risk for gastrointestinal complications related to use

of NSAIDs. A meta-analysis. Ann Intern Med 1991;115:787–96.

[18] Lanas A, Bajador E, Serrano P, et al. Nitrovasodilators, low-dose aspirin, other nonsteroidal

anti-inflammatory drugs, and the risk of upper gastrointestinal bleeding. N Engl J Med 2000;

343:834–9.

[19] Piper JM, Ray WA, Daugherty JR, et al. Corticosteroid use and peptic ulcer disease: role of

nonsteroidal anti-inflammatory drugs. Ann Intern Med 1991;114:735–40.

[20] Battistella M, Mamdami MM, Juurlink DN, et al. Risk of upper gastrointestinal hemorrhage in

warfarin users treated with nonselective NSAIDs or COX2 inhibitors. Arch Intern Med 2005;

165:189–92.

[21] Sorenson HT, Mellemkjaer L, Blot WJ, et al. Risk of upper gastrointestinal bleeding associated

with use of low-dose aspirin. Am J Gastroenterol 2000;95:2218–24.

[22] Kelly JP, Kaufman DW, Jurgelon JM, et al. Risk of aspirin-associated major upper-

gastrointestinal bleeding with enteric-coated or buffered product. Lancet 1996;348:1413–6.

[23] Weil J, Colin-Jones D, Langman M, et al. Prophylactic aspirin and risk of peptic ulcer bleeding.

BMJ 1995;311:259–60.

[24] Lanas A, Perez-Asa MA, Feu F, et al. A nationwide study of mortality associated with hospital

admission due to severe gastrointestinal events and those associated with nonsteroidal anti-

inflammatory drug use. Am J Gastroenterol 2005;100:1685–93.

[25] Kaiser J. Gender in the pharmacy: does it matter? Science 2005;308:1572–4.

[26] Cook NR, Lee IM, Gaziano JM, et al. Low-dose aspirin in the primary prevention of cancer:

the Women’s Health Study: a randomized controlled trial. JAMA 2005;294:47–55.

[27] Henry D, Dobson A, Turner C. Variability in the risk of major gastrointestinal complications

from nonaspirin nonsteroidal anti-inflammatory drugs. Gastroenterology 1993;105:1078–88.

[28] US Census Bureau. Gender: 2000. US Department of Commerce, Economics and Statistics

Administration. Available at: http://www.census.gov/prod/2001. Accessed August 10, 2005.

[29] Laine L. Approaches to nonsteroidal anti-inflammatory drug use in the high-risk patient.

Gastroenterology 2001;120:594–604.

[30] Higham J, Kang JY, Majeed A. Recent trends in admissions and mortality due to peptic ulcer in

England: increasing frequency of haemorrhage among older subjects. Gut 2002;50:460–4.

[31] Lipscomb GR, Campbell F, Rees WDW. The influence of age, gender, Helicobacter pylori and

smoking on gastric mucosal adaptation to non-steroidal anti-inflammatory drugs. Aliment

Pharmacol Ther 1997;11:907–12.

[32] Tarnawski A, Pai R, Tanigawa T, et al. Aging gastric mucosa has increased susceptibility to NSAIDs

injury—novel cellular and molecular mechanisms. Gastroenterology 2005;128(Suppl 2):A840.

[33] Lewis JD, Metz DC, Kimmel SE, et al. Pattern of use and risk of upper gastrointestinal toxicity

with over-the-counter nonaspirin nonsteroidal anti-inflammatory drugs (NANSAIDs). Gastro-

enterology 2005;128(Suppl 2):S1718.

[34] Dalton SO, Johansen C, Mellemkjaer L, et al. Use of selective serotonin reuptake inhibitors

and risk of gastrointestinal tract bleeding: a population-based cohort study. Arch Intern Med

2003;163:59–64.

[35] de Abajo FJ, Rodriguez LA, Montero D. Association between selective serotonin reuptake

inhibitors and upper gastrointestinal bleeding: population-based case-control study. BMJ 1999;

319:1106–9.

[36] Tata LJ, Fortun PJ, Hubbard RB, et al. Does concurrent prescription of selective serotonin

reuptake inhibitors and non-steroidal anti-inflammatory drugs substantially increase the risk of

upper gastrointestinal bleeding? Aliment Pharmacol Ther 2005;22:175–81.

[37] Takeuchi K, Tanaka A, Takahira Y, et al. Selective serotonin reuptake inhibitors (SSRIs)

aggravate antral ulcers induced by indomethacin in rat stomachs. Gastroenterology 2005;

128(Suppl 2):A337.

nonsteroidal anti-inflammatory drug injury 97

[38] Lancaster-Smith MJ, Jaderberg ME, Jackson DA. Ranitidine in the treatment of non-steroidal

anti-inflammatory drug associated gastric and duodenal ulcers. Gut 1991;32:252–5.

[39] Robinson MG, Griffin JW, Bowers J, et al. Effect of ranitidine gastroduodenal mucosal damage

induced by nonsteroidal anti-inflammatory drugs. Dig Dis Sci 1989;34:424–8.

[40] Silverstein FE, Graham DY, Senior JR, et al. Misoprostol reduces serious gastrointestinal

complications in patients with rheumatoid arthritis receiving nonsteroidal anti-inflammatory

drugs. Ann Intern Med 1995;123:241–9.

[41] Raskin JB, White RH, Jaszewski R, et al. Misoprostol and ranitidine in the prevention of

NSAID-induced ulcers: a prospective, double-blind, multicenter study. Am J Gastroenterol 1996;

91:223–7.

[42] Ekstrom P, Carling L, Wetterhus S, et al. Prevention of peptic ulcer and dyspeptic symptoms with

omeprazole in patients receiving continuous non-steroidal anti-inflammatory drug therapy. Scand

J Gastroenterol 1996;31:753–8.

[43] Cullen D, Bardhan KD, Eisner M, et al. Primary gastroduodenal prophylaxis with omeprazole

for non-steroidal anti-inflammatory drug users. Aliment Pharmacol Ther 1998;12:135–40.

[44] Hawkey CJ, Karrasch JA, Szczepanski L, et al. Omeprazole compared with misoprostol for

ulcers associated with nonsteroidal anti-inflammatory drugs. N Engl J Med 1998;338:727–34.

[45] Yeomans ND, Tulassay Z, Juhasz L, et al. A comparison of omeprazole with ranitidine for ulcers

associated with nonsteroidal anti-inflammatory drugs. N Engl J Med 1998;339:719–26.

[46] Yeomans N, Wilson I, Langstrom G, et al. Quality of life in chronic NSAID users: a comparison

on the effect of omeprazole and misoprostol. Scand J Rheumatol 2001;30:328–34.

[47] Hawkey C, Talley NJ, Yeomans ND, et al, and the NASA1 SPACE1 Study Group. Improvements

with esomeprazole in patients with upper gastrointestinal symptoms taking non-steroidal

anti-inflammatory drugs, including selective COX-2 inhibitors. Am J Gastroenterol 2005;100:

1028–36.

[48] Stupnicki T, Dietrich K, Gonzalez-Carro P, et al. Efficacy and tolerability of pantoprazole

compared with misoprostol for the prevention of NSAID-related gastrointestinal lesions and

symptoms in rheumatic patients. Digestion 2003;68:198–208.

[49] Graham DY, Agrawal M, Campbell DR, et al. Ulcer prevention in longterm users of non-

steroidal anti-inflammatory drugs. Arch Intern Med 2002;162:169–75.

[50] Yeomans N, Scheiman JM, Hawkey CJ, et al. An evidence-based analysis of esomeprazole

therapy versus placebo for the prevention of gastric or duodenal ulcers in at-risk continuous

NSAID users. Gastroenterology 2004;126(Suppl 2):W1278.

[51] Scheiman JM, Vakil N, Hawkey CJ, et al. Esomeprazole prevents gastric and duodenal ulcers

in at-risk patients on continuous non-selective or COX-2–selective NSAID therapy. Gastro-

enterology 2004;126(Suppl 2):A638.

[52] Sarosiek JT, Roeser SS, Connor M, et al. Restorative impact of rabeprazole on gastric mucus

and mucin production impairment during naproxen administration: its potential clinical sig-

nificance. Dig Dis Sci 2005;50:357–65.

[53] Spiegel BMR, Chiou C-F, Ofman JJ. Minimizing complications from nonsteroidal anti-

inflammatory drugs: cost-effectiveness of competing strategies in varying risk groups. Arthritis

Rheum 2005;53:185–97.

[54] Chan FK, Ching JY, Hung LC, et al. Clopidogrel versus aspirin and esomeprazole to prevent

recurrent ulcer bleeding. N Engl J Med 2005;352:238–44.