Few experimental data exist on the risk for gastroduodenal mucosal injury as a function of aspirin dose, and it is not known whether any clinically used dose of aspirin is free of risk for gastroduodenal mucosal damage. In a recent endoscopic study of patients with coronary artery disease, most of whom were receiving only 100 mg of aspirin per day, the prevalence of gastric erosions was higher than the prevalence of erosions in a historical control group [4].

The literature contains little information on the relation between aspirin dosage and suppression of gastroduodenal mucosal prostaglandin synthesis in humans. Because suppression of gastroduodenal mucosal prostaglandin synthesis appears to be one of the important mechanisms for mucosal damage by aspirin [5, 6], it would be useful to determine the threshold dose for gastroduodenal mucosal prostaglandin inhibition in humans because aspirin doses below this threshold may not damage the gastroduodenal mucosa.

Our goal was to determine the effects of a wide range of doses of aspirin (3 mg/d to 2600 mg/d) on gastric juice prostaglandins (PGE₂ and PGF₂ ) and on stomach mucosal injury as reflected by gastric juice hemoglobin and DNA, both of which are sensitive indicators of mucosal injury [7, 8]. Effects of various doses of aspirin were also related to their effects on gastric acid secretion and serum thromboxane B₂.

Methods

Top Methods Results Discussion Author & Article Info References

Healthy volunteers between the ages of 18 and 75 years were solicited through advertisement. All respondents had a screening history, physical examination, complete blood count, and serum electrolyte, creatinine, bilirubin, and liver chemistry tests. Volunteers were excluded for heartburn; abdominal pain; indigestion; bloating; fullness; nausea and vomiting; use (within the previous 2 weeks) of aspirin, nonsteroidal anti-inflammatory drugs, glucocorticoids, histamine-2-receptor antagonists, sucralfate, misoprostol, omeprazole, antacids, or Pepto-Bismol (Procter and Gamble, Cincinnati, Ohio); history of allergy to aspirin or topical tetracaine; a positive urine pregnancy test; history of gastric surgery, peptic ulcer, or other gastrointestinal disease; history of liver disease (including currently elevated liver chemistry tests); or history or presence of anemia, thrombocytopenia, leukopenia, gout, coagulation disorder, or renal insufficiency. Sixteen people (5 men and 11 women) met all entry criteria and gave written consent to participate as paid volunteers. Use of any medication other than study medication was not allowed during the study. The study was approved by the Human Studies Subcommittee at the Department of Veterans Affairs Medical Center, Dallas, Texas.

Part 1

Ten volunteers (four men and six women; mean age, 44.2 years; range, 19 to 73 years) were enrolled in part 1 of the study. One man was excluded for noncompliance. On 4 separate days at least 1 week apart, each of the other nine participants received placebo (calcium sulfate, 222 mg; cellulose, 97 mg); aspirin, 81 mg (Bayer Children's Aspirin, Glenbrook Laboratories; New York, New York); aspirin, 325 mg (one tablet); or aspirin, 650 mg (two 325-mg tablets). Medication was taken with each of three meals and at bedtime on the day before the study and at 0700 h on the study day. During the course of four visits, each of the nine volunteers took each of the four drug regimens in random order. Medications were dispensed by a research pharmacist so that the investigators were blinded to the study regimens. Although the medications differed in appearance, volunteers were blinded to exact ingredients of the medications and were instructed to refrain from discussing their study regimens with the investigators and technical assistants. We used the dosing schedule of the highest dose of aspirin (650 mg four times a day) in a previous study, which resulted in significant inhibition of gastric mucosal PGE₂ and PGF₂ concentrations [9].

Each volunteer was instructed to report to the research laboratory on the study day after an overnight fast. At 0800 h (1 hour after the fifth and final dose of medication), a venous blood sample was taken for determination of serum salicylate concentration. Then, a nasogastric tube was positioned fluoroscopically in the dependent portion of the gastric antrum. Volunteers were trained not to swallow their saliva and to collect saliva through dental suction catheters. Residual gastric secretion was aspirated with the volunteer in the supine position and the fluid was discarded. Beginning at 0830 h, gastric juice was collected by intermittent suction pump aspiration, and the samples collected were divided into aliquots and placed in separate containers every 15 minutes for the next 2 hours. In our laboratory, typical recovery of gastric juice with a nasogastric tube positioned fluoroscopically is greater than 90% [10].

Part 2

Eleven volunteers (three men and eight women; mean age, 45.2 years; range, 25 to 73 years) were enrolled in part 2 of the study. Two men were excluded for noncompliance. Four of the remaining nine volunteers (one man and three women) had participated in part 1 of the study. On five separate occasions at least 14 days apart, each volunteer received placebo (calcium sulfate, 222 mg; cellulose, 97 mg); aspirin, 3 mg; aspirin, 10 mg; aspirin, 30 mg; or aspirin, 81 mg for 8 days. Medications were prepared and packaged in identical green opaque gelatin capsules by the Pharmacy Service at Dallas Veterans Affairs Medical Center. Medications were dispensed by a research pharmacist so that the study was double-blinded. Volunteers were instructed to take the medication with breakfast for 7 days before the study and at 0600 h on the study day. During the course of five visits, each volunteer received each of the five drug regimens in random order. Each volunteer was instructed to report to the research laboratory after an overnight fast. At 0700 h (1 hour after the last dose of medication), venous blood samples were drawn for serum salicylate and serum thromboxane B₂ determinations. We chose to monitor serum thromboxane B₂ concentrations in part 2 because data from part 1 suggested that aspirin doses of 324 mg/d or less would not result in any detectable serum salicylate level and because the serum thromboxane B₂ level has been shown to be a clinically relevant and reliable marker of low-dose aspirin therapy [11]. After venipuncture, a nasogastric tube was positioned fluoroscopically in the gastric antrum, as described above. Beginning at 0730 h, gastric juice was collected by intermittent suction pump aspiration and samples collected were divided into aliquots and placed in separate containers every 15 minutes for the next 60 minutes. We chose to do this study for 60 minutes because results from part 1 indicated that no substantial fluctuations in gastric juice prostaglandin, hemoglobin, and DNA, or gastric acid outputs occurred throughout the 2-hour study period. The aspirin dosing schedule was chosen based on the following observations: 1) Thirty mg of aspirin daily has recently been shown to be as effective as 283 mg daily in the prevention of cerebrovascular events in patients with a transient ischemic attack or minor stroke [12]; 2) reduction of serum thromboxane B₂ by low-dose aspirin is cumulative on repeated daily dosing and has been shown to reach a steady state after 7 days [13]; and 3) serum thromboxane B₂ generally returns to normal by 14 days after cessation of aspirin therapy [13, 14].

Processing of Gastric Juice Samples and Determination of Gastric Acid Output

The volume of each 15-minute gastric juice sample was recorded to the nearest 0.2 mL, and indomethacin was added immediately to the sample after collection (final indomethacin concentration, 50 µmolars) to prevent further prostaglandin synthesis. The pH of each sample was determined by glass electrode. Acidity was derived from the pH [15] and expressed as mmol/L. Acid output was calculated by multiplying volume by acidity, expressed as mmol/h. To stabilize prostaglandins in the specimen, the sample was then titrated to pH 7.0 with 0.2-N NaOH and kept at 4 °C. Samples were processed the same day that they were collected and used in subsequent determinations of PGE₂, PGF₂ , hemoglobin, and DNA concentrations.

Determination of Gastric Juice PGE₂ and PGF₂ Output

Extraction of prostaglandins from gastric juice samples was done according to the method of Peskar and colleagues [16]. Aliquots of gastric juice that had been titrated to pH 7.0 were homogenized and centrifuged. The resultant supernatant was titrated to pH 3.0 with acetic acid. The acidified supernatant was extracted twice with four volumes of ethyl acetate. The organic extracts were pooled and stored at –70°C until radioimmunoassay. Before the assay, organic extracts were evaporated to dryness under a stream of nitrogen, and prostaglandins were reconstituted in a phosphate saline buffer. Recovery of prostaglandins was determined by adding 3500 dpm (Hydrogen-3)PGE₂ or (Hydrogen-3)PGF₂ to the specimen before the initial homogenization step and by counting an aliquot of the reconstituted specimen. Radioimmunoassay was done by incubating the reconstituted prostaglandins with the corresponding (Hydrogen-3)-prostaglandin and antiserum as described previously [17]. After incubation, bound counts were determined by liquid scintillation spectrometry. Standard curves using known amounts of PGE₂ and PGF₂ were used to determine the amounts of PGE₂ and PGF₂ present in the unknown sample. Prostaglandin E₂ and PGF₂ are the principal prostaglandins produced by gastric mucosa and released in gastric juice in humans [5, 16].

Determination of Gastric Juice Hemoglobin and DNA

Hemoglobin was determined using a kit (Sigma; St. Louis, Missouri). Aliquots of gastric juice that had been titrated to pH 7.0 were sonicated to disrupt cell membranes and release DNA. The DNA concentration was determined in part 1 by a method that uses the enhancement of fluorescence seen when bisbenzamidazole (Sigma) binds to DNA [18]. The DNA concentration in part 2 was determined by a method using diphenylamine [19] because production of bisbenzamidazole had been discontinued by the manufacturer. Salmon testes DNA (Sigma) was used as a standard in parts 1 and 2.

Determination of Serum Salicylate and Serum Thromboxane B₂ Concentrations

Serum salicylate was measured using a kit (Sigma). Values of 0.145 mmol/L or less are below the limit of sensitivity established by the manufacturer and are considered "negative" (undetectable).

The serum thromboxane B₂ level was determined from blood collected by venipuncture into glass tubes according to the method described by Kallmann and associates [11]. A blood sample of 10 mL was allowed to clot for 45 minutes at 37 °C and then for 2 hours at 20 °C. At the end of the 2.75-h incubation period, indomethacin was added to the sample (final concentration, 50 µmolars). Serum was separated by centrifugation (2000 g, 10 minutes, 20 °C) and stored at –70°C until a radioimmunoassay was done. Serum thromboxane B₂ was measured using a kit (Amersham; Arlington Heights, Illinois).

Statistical Analysis

Differences in mean results between placebo and aspirin are expressed as the average difference with 95% confidence intervals. Differences in mean results with various treatment regimens were also compared using an analysis of variance followed by paired t-tests [20, 21]. Dose-response curves were generated with the aid of a computer curve-fitting program (Sigmaplot, Version 5.0, Jandel Scientific; San Francisco, California) and differences between various curves were compared using analysis of variance. Statistical analyses were done using Instat, Version 1.12a (GraphPAD Software; San Diego, California). Probability values 0.05 were considered statistically significant.

Results

Top Methods Results Discussion Author & Article Info References

Part 1

Mean serum salicylate levels were below 0.145 mmol/L (2 mg/dL) in volunteers who took either placebo or 324 mg of aspirin per day in four divided doses, whereas mean serum salicylate levels of volunteers receiving daily aspirin doses of 1300 mg and 2600 mg (in four divided doses) were 0.30 mmol/L (4.1 mg/dL) and 0.58 mmol/L (8.1 mg/dL) (both P <0.001 compared with placebo).

All three aspirin doses significantly reduced mean gastric juice PGE₂ and PGF_{2 concentrations} and outputs to approximately the same degree. As indicated in Figure 1 (top left), mean reductions of gastric juice PGE₂ output by daily aspirin doses of 324 mg, 1300 mg, and 2600 mg per day were 48% (95% CI, 8% to 88%), 62% (CI, 25% to 99%) and 54% (CI, 22% to 87%), respectively. Mean reductions of gastric juice PGF₂ output were 40% (CI, 16% to 64%), 49% (CI, 28% to 70%), and 37% (CI, 20% to 53%), respectively (Figure 1, top right).

View larger version (22K): [in this window] [in a new window]   Figure 1. Effects of aspirin dose on gastric juice outputs. Results are expressed as a mean percent change from placebo for nine volunteers, with 95% CIs. Mean (±SE) outputs for the zero dose (placebo) were 108.9 ±25.0 ng/h (PGE2); 56.1 ±9.4 ng/h (PGF2 ); 16.7 ±5.6 mg/h (hemoglobin); and 436.4 ±130.8 µg/h (DNA). * P < 0.05 compared with placebo.

Part 2

Mean serum salicylate levels determined 1 hour after the last aspirin dose were, as expected, below 0.145 mmol/L (2 mg/dL) in all treatment groups. All aspirin doses, including 3 mg/d, significantly reduced serum thromboxane B₂, and the inhibitory effect was clearly dose related (Figure 2, top left). In contrast, only aspirin at a dose of 30 mg/d significantly reduced gastric juice PGE₂ concentration and PGE₂ output (Figure 2, top right), although there was a trend toward lower gastric PGE₂ outputs with all other aspirin doses. No significant changes in gastric juice PGF₂ concentration and PGF₂ output (Figure 2, top right) were noted at any aspirin doses in part 2. Moreover, no significant changes in gastric juice hemoglobin or DNA concentrations or outputs were observed (Figure 2, bottom left and right). Gastric juice volume, acidity, and acid output were not significantly affected by any aspirin dose.

View larger version (23K): [in this window] [in a new window]   Figure 2. Effects of aspirin dose on serum thromboxane B2 and gastric juice outputs. Results are expressed as a mean percent change from placebo for nine volunteers, with 95% CIs. Mean (±SE) outputs for the zero dose (placebo) were 700 ±60.9 ng/mL (serum thromboxane B2 [TxB2]); 78.4 ±16.9 ng/h (PGE2); 30.9 ±8.5 ng/h (PGF2 ); 7.4 ±1.4 mg/h (hemoglobin); and 885.6 ±137.2 µg/h (DNA). * P < 0.05 compared with placebo.

Figure 3 compares the dose-related inhibitory effects of aspirin on serum thromboxane B₂ level and gastric juice PGE₂ output. The inhibitory effect of aspirin on serum thromboxane B₂ was significantly greater than the suppressive effect of aspirin on gastric juice PGE₂ output at all tested doses (P = 0.05, P = 0.01, P = 0.004, and P = 0.004 with 3, 10, 30, and 81 mg/d, respectively). The aspirin dose that reduced serum thromboxane B₂ by 50% was approximately 3 mg/d. In contrast, a 50% inhibition of gastric PGE₂ output was achieved with an aspirin dose of approximately 30 mg/d, and higher doses did not result in further inhibition.

View larger version (15K): [in this window] [in a new window]   Figure 3. Dose-related inhibitory effects of aspirin on serum thromboxane B2 and gastric juice PGE2 outputs. Percent change was calculated for each participant compared with his or her own control value. Each data point represents the mean percent change with 95% CI. = thromboxane B2 (TxB2) data from part 2; \#9679; =gastric juice prostaglandin (PGE2) data from part 1; = PGE2 data from part 2. * P < 0.05 compared with the PGE2 curve.

Discussion

Top Methods Results Discussion Author & Article Info References

In part 1 of our study, similar inhibition of gastric juice prostaglandin output was observed with daily aspirin doses ranging from 324 to 2600 mg, whereas significant stomach mucosal damage, as estimated by an increase in gastric juice hemoglobin output, occurred only with the highest aspirin dose (2600 mg/d). Our results indicate that higher doses of aspirin may acutely damage the gastric mucosa by a prostaglandin-independent mechanism because daily aspirin doses of 324, 1300, or 2600 mg inhibited prostaglandin outputs by around 50%, yet only the last dose caused significant acute gastric mucosal injury as reflected by hemoglobin output.

A recent study from the Netherlands indicated that 30 mg of aspirin daily was as effective as a 283-mg dose in the prevention of cerebrovascular events in patients with a transient ischemic attack or minor stroke and that the 30-mg dose was associated with significantly fewer minor bleeding complications [12]. These observations and data from part 1 of our study prompted us to do part 2, which was designed to investigate the effects of very low doses of aspirin on serum thromboxane B₂, gastric prostaglandin output, and mucosal injury. Our results showed that significant inhibition of serum thromboxane B₂ occurred at extremely low doses of aspirin, consistent with results from other studies [11, 13, 14]. On the other hand, significant inhibition of gastric PGE₂ output was not noted at aspirin doses below 30 mg/d. Because PGI₂, another gastric prostaglandin [22], was not measured, an inhibitory effect of low-dose aspirin on this prostanoid cannot be excluded.

Why the stomach appears to be more resistant than the platelet to aspirin-induced inhibition of arachidonic acid metabolism is uncertain. Selective sparing of cyclooxygenase activity in the stomach relative to the platelet may reflect the capability of gastric mucosal cells, which are nucleated, to synthesize new enzyme de novo [23]. Taken together, our observations show the hierarchical effects of aspirin on serum thromboxane B₂ (most sensitive), gastric PGE₂ output (less sensitive), and gastric mucosal injury (least sensitive). Our data may also provide an explanation for previous observations that gastric mucosal PG suppression by aspirin or indomethacin may not always be accompanied by gastric mucosal injury [6, 17, 24].

Although almost complete (>98%) reduction of serum thromboxane B₂ was attained with an aspirin dose of 30 mg/d, only a 50% inhibition of gastric PGE₂ output was achieved with this dose, and higher aspirin doses did not result in further PGE₂ inhibition. Moreover, daily aspirin doses ranging from 324 mg to 2600 mg resulted in an approximately 50% inhibition of gastric PGF₂ output. Our data are consistent with previous observations reported by Cohen and MacDonald [5] that a daily aspirin dose of 2600 mg (for 5 days) resulted in an approximately 60% reduction in PGE₂ content in both gastric juice and gastric mucosal specimens. Similarly, Child and associates [25] reported that a daily aspirin dose of 3300 mg (for 5 days) resulted in a 40% reduction in gastric juice PGE output. We speculate that this incomplete and unequal suppression of PGE₂ or PGF₂ output by aspirin is caused by the differential effects of aspirin on various types of gastric mucosal cells that synthesize these prostaglandins, as well as their relative contribution to the overall prostaglandin output in gastric juice. Further studies using isolated human gastric mucosal cells are needed to ascertain whether the PGF₂ -producing gastric mucosal cells are more resistant to low-dose aspirin than the PGE₂-producing cells, as suggested by our data from part 2 (see Figure 2).

In parts 1 and 2, we did not observe any significant changes in gastric acidity or acid output in any treatment group. Our data are consistent with previous observations that oral aspirin had no significant effects on gastric acidity and gastric acid output in healthy human volunteers [25, 26].

Although we attempted to enroll roughly equal numbers of male and female volunteers in our study, considerably more female than male volunteers participated in both parts, mainly because several male participants had to be excluded from the study because of noncompliance. However, we believe that the female predominance in our study group did not affect our findings and conclusions because we have previously shown that sex has no significant effect on gastric mucosal prostaglandin concentrations [27] or on the severity of aspirin-induced acute gastric mucosal injury as documented endoscopically [9].

This dose-response study presents the first experimental evidence that aspirin can significantly reduce serum thromboxane B₂ at doses that are significantly lower than the threshold dose for significant gastric prostaglandin inhibition in humans and that do not cause any acute gastric mucosal injury in healthy volunteers. For instance, 10 mg/d of aspirin reduced serum thromboxane B₂ by 75% without significantly suppressing gastric PGE₂ or PGF₂ outputs or inducing any acute gastric mucosal injury. Because thromboxane B₂ in serum is mostly derived from platelets [11], the clinical implication of our results is that a low dose of aspirin, such as 3 mg/d or 10 mg/d, may selectively inhibit thromboxane-dependent platelet functions and reduce coronary and cerebral thrombotic events without substantially altering gastric mucosal prostaglandin metabolism or inducing gastric ulcerations. The absence of significant gastric mucosal damage with acute administration (8 days) of aspirin in doses of 10 mg/d or less is reassuring but does not guarantee that mucosal damage would not occur with a longer duration of therapy at the same doses. Therefore, prospective, placebo-controlled trials are needed to ascertain whether long-term, very-low-dose aspirin prophylactic therapy is effective in preventing thrombotic events while also associated with a low incidence of gastrointestinal ulcers and their frequently life-threatening complications.

This work was presented in part at the 93rd and 94th annual meetings of the American Gastroenterological Association, 10 to 13 May 1992, in San Francisco, California, and 17 to 19 May 1993, in Boston, Massachusetts, and was published in abstract form (Gastroenterology. 1992; 102:A53; and 1993; 104:A131).