If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Background & Aims: Patients who regularly take nonsteroidal anti-inflammatory drugs (NSAIDs) have an increased risk for small-intestinal mucosal ulceration and bleeding, which may present as anemia of undetermined gastrointestinal origin or protein loss. The prevalence and severity of small-intestinal lesions remains unclear. Our aim was to assess the frequency of NSAID-induced small-bowel injury among chronic NSAID users. Methods: Ambulatory patients with various types of arthritides who took NSAIDs daily (>3 mo duration) or took either acetaminophen alone or nothing were enrolled in the study. All patients fasted overnight and underwent wireless video capsule endoscopy. Two investigators, blind to therapy, reviewed each video beginning after the pylorus. Lesions were scored as normal, red spots, small erosions, large erosions, or ulcers. An ulcer was defined as a larger lesion with apparent depth and a definite rim. Results: Forty-one patients, 36 men and 5 women, ages ranging from 22 to 66 years (mean age, 49.8 y) were analyzed including 21 chronic NSAID users and 20 control patients. Small-bowel injury was seen in 71% of NSAID users compared with 10% of controls (P < .001). Injury was mild (few or no erosions, absence of large erosions/ulcers) in 10 NSAID users compared with 2 controls. Five NSAID users had major (>4 erosions or large ulcers/ulcers) damage compared with none in the control group. There were no complications or problems with the capsule endoscopy procedure. Conclusions: Endoscopically evident small-intestinal mucosal injury is very common among chronic NSAID users. The role of endoscopically evident injury in unexplained iron-deficiency anemia and hypoalbuminemia among chronic NSAID users remains undetermined.
Nonsteroidal anti-inflammatory drug (NSAID) therapy has great benefits in terms of decreases in inflammation and pain, but these benefits are offset by the occurrence of potentially life-threatening complications related to NSAID-induced gastrointestinal injury. Until recently, the focus has been on NSAID damage to the esophagus, stomach, and duodenum because these are the sites of major morbidity and mortality.
NSAIDs also can cause a variety of functional and structural abnormalities in the small intestine such as increased intestinal permeability, intestinal inflammation, protein loss, blood loss, ulcerations, perforation, diaphragm-like strictures, and ileal dysfunction.
The general lack of correlation between intestinal blood loss and the presence of drug-associated gastroduodenal lesions led to the realization that NSAID-induced small-intestinal mucosal ulceration and bleeding were clinically important, typically presenting as iron-deficiency anemia or hypoalbuminemia.
The small intestine is largely inaccessible to direct observation. The largest study to examine the prevalence and severity of NSAID-induced small-intestinal lesions was a postmortem examination of the stomach, duodenum, and small intestine of 713 patients
; 249 had NSAIDs prescribed during the 6 months before death and their findings were compared with the 464 patients who did not. Nonspecific small-intestinal ulcerations were found in 8.4% of the users of NSAIDs compared with .6% of the nonusers (P < .001). Of interest, 3 patients who were long-term users of NSAIDs were found to have died of perforated nonspecific small-intestinal ulcers.
A number of studies evaluated indirect parameters of mucosal damage by NSAID intestinal permeability or inflammation.
Taken together, these data are consistent with the notion that the effect of intestinal permeability is primarily a local intestinal event and that COX-1 inhibition plays a role.
The purpose of this study was to assess the use of video capsule endoscopy for detecting NSAID-induced small-intestinal injury in arthritis patients who chronically used NSAIDs compared with a group of patients who did not.
This was an open-label, endoscopist-blind, prevalence study of small-bowel injury associated with chronic NSAID use. Entry criteria included generally healthy men or women, ages 18–70 years, with at least a 3-month history of daily NSAID use for osteoarthritis, rheumatoid arthritis, or nonspecific arthritis and were nonusers of NSAIDs (control subjects). None had anemia or hypoalbuminemia. Patients had to be in good general health. Those with evidence suggesting intermittent or partial obstruction of the small intestine, swallowing disorders, a history of poor gastric emptying, or prior gastric surgery were excluded. Regular use of antisecretory drugs such as H2-receptor antagonists or proton pump inhibitors was permitted. The use of misoprostol caused exclusion whereas aspirin use (≤325 mg/day) was allowed among the chronic NSAID-user group. No NSAIDs were allowed on the day of the procedure such that the last dose permitted was the evening before the procedure.
Potential patients with a diagnosis of osteoarthritis or rheumatoid arthritis who were users of nonselective NSAIDs were identified from the general medicine clinics of the Veteran’s Administration Medical Center from routine review of the pharmacy distribution record provided during each encounter. Those patients meeting the entry criteria were invited to participate. Controls were chosen because they had some form of arthritis, usually nonspecific degenerative joint disease of a lower extremity or spine, often as a result of chronic physical activity (warehouse workers and laborers) or vigorous exercise (gymnastics and runners), and because they did not use NSAIDs.
Patients were instructed to fast for at least 8 hours (overnight) before arriving at the facility. Patients using NSAIDs were instructed to take their morning dose with 8 oz water by mouth at least 8 hours before the procedure and, if applicable, patients skipped their morning dose on the day of the procedure. After set-up of the data collecting system and attachment of the recorder to the patient, the capsule was swallowed. The patients remained in the endoscopy unit for 1 hour and then were allowed to move freely about the hospital. The monitoring period was approximately 7 hours. Ingestion of water or plain tea was permitted after 2 hours and a small lunch (eg, a turkey sandwich, cookie, and 6 oz apple juice) was provided 4 hours into the procedure. Before discharge, the patients were instructed to contact the study physician if any abdominal pain, vomiting, or other gastrointestinal disturbance occurred.
Capsule device and equipment
The video capsule endoscopy system (Given Diagnostic Imaging System; Given Imaging LTD, Yoqneam, Israel) is 26 mm long, 11 mm in diameter, weighs 3.45 g, has a field of vision of 140°, and an approximate magnification power of 1:8. It is powered by 2 silver oxide batteries that last approximately 8 hours, during which time it records approximately 55,000 images.
Demographics, health information, and all medications used were recorded. Digital images obtained from individual patient procedures were downloaded to CD-ROM discs, in duplicate, and labeled with patient initials and the sequential study enrollment serial number. One unblinded investigator performed a preliminary review of all images to annotate (thumb-nails) the pyloric valve at the time of passage to the duodenum and the ileocecal valve immediately before passage to the cecum or terminal ileum region where the video capsule endoscope first became stagnant. Two independent investigators who remained blinded to the patient group (NSAID user or control) separately reviewed each of the procedures for intestinal injury beginning with the time of passage through the pylorus to passage to the cecum. The blinded reviewers did not examine the stomach before examination of the small bowel to prevent bias from knowing if gastric lesions were present. Descriptive data and the location of lesions seen were recorded as thumb-nail photographs. After the independent review was complete, the findings were reviewed in conference by both blinded investigators to reach a consensus regarding the degree of injury. All findings were recorded. Images seen on capsule endoscopy generally are magnified and usually are seen close-up, introducing difficulty in judging the size. The descriptions of the findings were scored based on prior extensive experience with NSAID-induced damage to the stomach and proximal duodenum. The categories were normal, red spots, small erosions, large erosions, or ulcers. No mucosal webs or strictures were seen. Mucosal folds were not characterized as other than present or absent. A red spot was defined as a red or crimson area of mucosa with preservation of villous architecture. With capsule endoscopy it is difficult or impossible to distinguish large erosions from ulcers because capsule endoscopy does not allow repeated observation, assessment of the effects of motility, or viewing from different angles. Nonetheless, in the stomach and duodenal bulb both the number and size are thought to be important predictors of risk. A small erosion was defined as a circumscribed area of mucosal disruption denuded of villi with or without exudates or red color that involved, at most, the diameter equivalent of a valvulae conniventes (Figure 1). Large erosions were defined as circumscribed breaks in the mucosa larger than the equivalent diameter of a valvulae conniventes. Ulcers were defined as large erosions with a central area with exudates, typically with a surrounding border of elevated mucosa (eg, apparent depth), producing a target lesion or coral polyp appearance (Figure 2). For scoring purposes, ulcers were combined with large erosions into the category of large erosion/ulcer. Lesions were scored but the study code was not broken until after all analyses were completed.
Video capsule endoscopy does not allow for a good examination of the stomach or duodenal bulb and thus it was not possible to attempt any correlations between gastric and small-bowel findings. Statistical analyses were performed using SigmaStat 3.02 (SPSS, Chicago, IL). Parametric data were analyzed using the Student t test and proportions were analyzed by χ2 or Fisher exact test, whichever was appropriate.
The sample size was chosen prospectively based on the predictions from the literature showing 8.4% (95% confidence interval [CI], 5%–12.6%), with erosions in an autopsy study
(95% CI, 21%–73%). The primary goals of the study were to determine whether video endoscopy was able to detect NSAID-associated small-bowel injury and to obtain a reasonable ballpark estimate of the frequency. We assumed that the frequency would be higher than that seen at autopsy because autolysis and other changes would not be present in vivo. We based the sample size on the 95% CI for a 20% prevalence to ensure that if the prevalence was as low as 20% we would find at least 1 patient with a lesion. An identically sized control group was chosen to ensure that the findings were not similar in those not taking NSAIDs. It was accepted that a significant difference between NSAID users and controls would be achieved only if the prevalence of small-bowel changes were greater than 50% among NSAID users and rare among controls. Nonetheless, the study was planned to provide a good estimate with a 95% CI for NSAID-induced damage and for the prevalence of similar lesions among controls.
A scoring system was devised based in part on prior experience with NSAID-induced gastric mucosal damage with consideration of the fact that the image is magnified 8-fold in capsule endoscopy and it is impossible to determine accurately or precisely the size of the lesions, making any scoring system somewhat subjective or anecdotal. Nonetheless, as defined earlier, it is possible to distinguish small vs. large or not small. Initial scores then were grouped to yield 3 categories: normal = 0, mild damage = 1–2, and major damage = 3–4 (Table 1). The grouping was based on previous work in NSAID-induced gastric damage that more erosions are worse than few erosions and that large erosions and ulcers cannot be distinguished reliably.
Table 1Five-Point Scoring System for Endoscopic Lesions With Capsule Endoscopy
Forty-three patients, 38 men and 5 women, ages ranging from 22 to 66 years (mean age, 49.8 y) were enrolled. The control group consisted of 17 men and 3 women vs. 19 men and 2 women in the NSAID-user group. The mean age was 44.5 years for controls vs. 55.2 for NSAID users (P = .001). The types of arthritis were as follows: osteoarthritis-degenerative joint disease, 20 vs. 18; rheumatoid arthritis, 0 vs. 1; gouty arthritis, 0 vs. 3; periostitis, 1 vs. 0; for controls vs. NSAID users, respectively. One control was replaced for a protocol violation (ie, use of low-dose aspirin), which was precluded in the control group. The replacement was performed before analysis of the video records and the data were not included. One control was excluded because of the finding of extensive celiac disease. Although no erosions or ulcers were seen, it is unknown whether the mucosa in this disease is similar to normal mucosa in terms of the propensity toward NSAID damage. This patient was not replaced. The final sample size included 21 chronic NSAID users and 20 control patients.
Drug usage by the controls consisted of acetaminophen in 12 patients. No drugs thought to damage the intestinal mucosa were used. Among the NSAID users the drugs used included naproxen (n = 13), ibuprofen (n = 6), indomethacin (n = 1), and oxprozocin plus aspirin (n = 1); and combined therapy included naproxen plus aspirin (n = 6), ibuprofen plus piroxicam plus aspirin (n =1), and 1 patient taking naproxen also received methotrexate, prednisone, and sulfasalazine. The daily doses were as follows: naproxen 500 mg (n = 5), 750 mg (n = 1), 1000 mg (n = 7); ibuprofen 1200 mg (n = 3), 1600 mg (n = 1), 2400 mg (n = 2); indomethacin 75 mg (n = 1); piroxicam 20 mg (n = 1); oxyprozocin 600 mg (n = 1); and aspirin 325 mg (n = 8). No other drugs thought to damage the intestinal mucosa were used.
Small-bowel injury was seen in 71% of NSAID users compared with 10% of controls (P < .001) (Table 2). Injury was mild in 10 NSAID users compared with 2 controls. The most severe damage among the controls was 1 small erosion. Five NSAID users had more impressive damage compared with none in the control group. The most severe damage was the presence of ulcers (Figure 1). There were no complications or problems with the capsule endoscopy procedure.
suggesting that damage to the small intestine is a common event. Study of the problem proved difficult. Barium radiographic examination of the small-bowel mucosa is incapable of detecting flat lesions and generally was normal despite occult bleeding and protein loss. The introduction of fiberoptic enteroscopy offered the promise of a practical method to study the small-bowel mucosa but the Sonde instrument proved cumbersome to use and essentially was abandoned when improved push enteroscopes became available. Even these improved instruments proved difficult, and sometimes dangerous or they may not reach the site of the lesions such that their use largely has been restricted to clinical situations in which the risk-benefit calculation is decidedly on the side of benefit. Thus, the available data regarding the types and location of damage largely has been limited to autopsy studies and clinical cases with adverse outcomes.
Video capsule endoscopy offered the promise of a breakthrough for the evaluation of small-bowel disease and current evidence suggests that it has an important role in the evaluation of patients with bleeding from the small intestine
As is the case in most newly introduced technologies, the true indications and recognition of the limitations still is evolving.
This trial used video capsule endoscopy to examine the effect of NSAIDs on the small intestine in an organized and blinded manner. We found significant mucosal damage in the majority of patients taking NSAIDs chronically. Ulcers and erosions were markedly more common than had been suggested by the previous autopsy study.
Of interest, the percentage with mucosal damage was approximately the same percentage reported in the studies of intestinal inflammation and comparison of the results of the 2 techniques in the same patients seems warranted. It is also of interest that Morris et al.
reported similar results when they performed Sonde enteroscopy to examine the small intestine in a highly selected group of 15 patients with rheumatoid arthritis presenting with chronic occult gastrointestinal blood loss and iron-deficiency anemia. All of their patients had normal upper-gastrointestinal endoscopy and colonoscopy. They assigned NSAID-associated small-bowel lesions into 3 categories: (1) red spots, defined as demarcated, usually circular, 1–3-mm areas of crimson mucosa with preservation of villous pattern; (2) ulceration/erosion, defined as a red spot with clear loss of villi across the area of the lesion, <5 mm in diameter with a definite edge; and (3) ulcers, defined as a penetrating lesion of mucosa with a diameter of >5 mm.
Seven (47%) had small-bowel ulceration and an additional 3 (20%) had red spots. These results (66% with damage) are remarkably similar to those presented here and to the finding of studies using intestinal inflammation as the endpoint, suggesting that studies of intestinal inflammation may be a surrogate for visible damage.
Video capsule endoscopy also showed that NSAID damage is more frequent and extensive than suggested by studies of NSAID-associated small-bowel injury shown by ileoscopy performed at the time of colonoscopy
The high frequency of small-bowel erosions found in this study also is similar to that reported in those taking traditional NSAIDs in a recent comparison of traditional NSAIDs and COX-2–selective inhibitors,
showing that erosions are common. In that study, patients were screened by capsule findings before NSAID therapy and after NSAID therapy. This study (eg, the 95% CI of the prevalences) provides the data needed to estimate the prevalence of damage for subsequent studies. Subsequent studies are needed to investigate whether there is a pattern of damage reliably associated with outcomes such as unexplained iron-deficiency anemia or hypoalbuminemia among chronic NSAID users, as well as the effect of different NSAIDs including aspirin, especially enteric-coated aspirin.
Hospitalization and mortality rates from peptic ulcer disease and GI bleeding in the 1990s.
Research supported in part by the Office of Research and Development Medical Research Service Department of Veterans Affairs, Hilda Schwartz G.I. Endowment, and by Public Health Service grants DK53659 and DK56338, which funds the Texas Gulf Coast Digestive Diseases Center.
Video capsules and the associated computer workstation were supplied by Given Imaging, LTD. This was an investigator-initiated protocol. Given Imaging had no input into design or analysis.