Allan R. Cooke

The role of acid in the pathogenesis of aspirin-induced gastrointestinal erosions and hemorrhage.

Aspirin is capable of damaging the gastric mucosa and this can be detected by a variety of technics. The damage results in exfoliation of the gastric mucosal cells, erosion of the gastric mucosa, and loss of blood from the stomach either as occult or overt bleeding. Salicylates in doses of about 3 g daily will cause gastrointestinal blood loss of about 2 to 6 ml daily in 60 to 70~ of normal subjects. This blood loss is usually not clinically significant since the great majority of patients with rheumatoid arthritis can to lera te aspir in w i thou t any increase in anemia (1). However, an occasional subject receiving similar doses of salicylate loses much larger amounts of blood and some may even bleed overtly. The blood loss is such (10 to 100 ml daily) that aspirin is an occasional cause of anemia. There is also epidemiologic evidence associating aspirin ingestion with hematemesis and melena, but this evidence remains controversial (2-4). It has been suggested that acid does not play a vital role in precipitating gastric mucosal damage (5, 6) and that occult bleeding and overt bleeding associated with aspirin are unrelated phenomena (7). It is the view of this author that all the conflicting studies of aspirin effects on the stomach can be resolved to fit a simple hypothesis, the basis of which involves the presence of acid.

SHAM FEEDING AND PANCREATIC SECRETION IN THE DOG

Summary Sham feeding caused a marked increase in pancreatic protein output in dogs with permanent pancreatic and gastric fistulas. This effect could not be attributed to secondary pancreatic stimulation by passage of gastric juice into the duodenum, as it could be shown that the gastric fistula effectively prevented duodenal acidification during a secretory response to gastrin or histamine. The pancreatic response to sham feeding was inhibited by acidification of an innervated pouch of the pyloric gland area in a second group of dogs. This suggests that part of the pancreatic response to sham feeding is mediated by the vagal release of gastrin from the pyloric gland area.

LOCALIZATION OF RECEPTORS INHIBITING GASTRIC EMPTYING IN THE GUT

Five dogs were prepared, each with three fistulas; one fistula was placed in the stomach, the other in the first part of the duodenum, 5 cm from the pylorus, and the third at the junction of the duodenum and the jejunum. By combinations of test meals and perfusions, either a part of or the whole duodenum or jejunum was tested with 550 to 850-milliosmolar solutions (glucose, glycine, mannitol, or sodium chloride), 100 mm HCl, 10 or 20 mm sodium oleate and 40 mm tryptophan for their effect on gastric emptying. Phenol red was used as a nonabsorbable marker. Gastric emptying was not slowed by perfusion of the whole duodenum with the hypertonic solutions of glucose, glycine, mannitol and sodium chloride, or with 10 or 20 mm sodium oleate. In contrast, significant slowing of gastric emptying occurred with jejunal perfusion of the hypertonic solutions and sodium oleate. Jejunal perfusion of 100 mm HCl slowed gastric emptying, but perfusion of the second to fourth parts of the duodenum did not. Tryptophan slowed gastric emptying when allowed to perfuse the first 5 cm of the duodenum, the second to fourth parts of the duodenum, and also the jejunum. The findings from this study indicate that the jejunum has receptors for inhibition of gastric emptying responsive to tryptophan, acid, fat, and hypertonic solutions. In contrast, the duodenum has receptors responsive to tryptophan.

Effect of First Part of Duodenum on Gastric Emptying in Dogs: Response to Acid, Fat, Glucose, and Neural Blockade

Five dogs were prepared, each with a gastric and duodenal fistula (5 cm distal to the pylorus), to study the inhibitory role of the first 5 cm of the duodenum on gastric emptying. The basic design of the experiments was to instill the test meal (300 ml at 37 degrees C, containing phenol red 40 mg 1(-1)) into the stomach and collect it at 1- or 2-min intervals for 10 or 20 min from the duodenal fistula. As the test meal emptied from the stomach it bathed the first 5 cm of duodenum and thus stimulated the appropriate receptor. A Foley catheter with an inflated balloon prevented passage into the second part of the duodenum. Test meals of hypertonic glucose (15%, 865 milliosmoles kg-1) or 20 and 80 mM of sodium oleate emptied at the same rate as water when allowed to bathe the first 5 cm of duodenum, whereas test meals of 100 mM of HCl were slowed. In further studies using neural blocking agents, the emptying of water meals was slowed with subcutaneous atropine sulfate (0.03 and 0.15 mg kg-1), intravenous hexamethonium chloride (10 mg kg-1), and norepinephrine bitartrate (0.04 mg kg-1). The emptying of 100 to 120 mM HCl meals was slowed by subcutaneous atropine sulfate (0.03 and 0.15 mg kg-1), intravenous norepinephrine bitartrate (0.04 mg kg-1), and the intravenous alpha-receptor blocking agents phenoxybenzamine HCl (2 mg kg-1) and phentolamine (2 mg kg-1), was unaffected by intravenous hexamethonium chloide (10 mg kg-1), and was unchanged (1.0 mg kg-1) or slightly slowed (2.0 mg kg-1) by the beta-receptor blocker propranolol. In contrast, acid test meals were emptied at the same rate as water when treated with intravenous guanethidine monosulfate (2 mg kg-1) or intramuscular reserpine (1 mg kg-1), indicating that the acid inhibition was mediated by an adrenergic mechanism. The emptying of water meals was unchanged by these two drugs. The authors suggest that the first 5 cm of duodenum contain receptors for inhibition of emptying of acid but not for fat or hypertonic glucose. Furthermore, the neural blocking studies indicate that the inhibitory effect of acid in the first part of the duodenum is an adrenergic mechanism which appears to be neither alpha nor beta-receptor-mediated.

The role of the mucosal barrier in drug-induced gastric ulceration and erosions

There are seven drugs usually stated to be ulcerogenic in man and at least one general textbook of medicine implicates a number of them (1). These drugs are adrenal corticosteroids, aspirin, phenylbutazone, indomethacin, ethanol, caffeine, and tobacco (smoking). The types of data upon which these conclusions are based follow three lines of evidence. First, the ability of these drugs to produce experimental ulcers or erosions; all of these drugs can cause ulcers when given to animals (2-7). Second, a cause-andeffect relationship in man, ie, epidemiological evidence; the epidemiological evidence is very weak except for aspirin and smoking. Third, a mechanism suspected of participating in the pathogenesis; the pathogenesis of drug ulceration is not fully understood but aspirin may be the only one with a body of data to support its ulcerogenic effect. The purpose of this paper is to examine the epidemiological evidence as well as possible mechanisms by which drugs may be associated with gastric damage. Reviews of various aspects of this subject have been published in the past few years (8-12). Thus, attention will be mainly directed to certain aspects of the literature.

Control of gastric emptying by amino acids.

Studies were made of gastric emptying of amino acids in 6 healthy subjects using test meals (350 ml) containing phenol red and the appropriate amino acid. Gastric contents were aspirated at 20 min and samples estimated for phenol red concentration and titratable acidity. The amino acids were glycine, β-alanine, L-leucine, L-lysine, and DLmethionine and the concentrations were 0.1 m, 0.2 m, 0.5 m, and 1.0 m. It was found that all amino acids in concentrations of 0.2 m delayed emptying significantly when compared with test meals containing water and phenol red only. The delay in emptying was related to the molar concentration of the amino acid in the test meal and unrelated to the molecular weight of the amino acid or to the amount of acid secreted. The results of these studies indicate, that gastric emptying of amino acids is related to their osmolar concentration.

Studies of Anti-Inflammatory Drugs and Aliphatic Alcohols on Antral Mucosa

Studies were made in 4 dogs with antral pouches of transmucosal potential difference (PD), and Na+, H+, K+, and Cl− fluxes, in response to aspirin, phenylbutazone, indomethacin, and cortisone acetate, as well as the aliphatic alcohols of carbon chain lengths C1 to C4. Potential difference was measured using 3% agar-saturated KCl electrodes connected to a potentiometer. Net fluxes of electrolytes were measured after instillation of 15 to 20 ml of a solution of 100 mm HCl + 54 mm NaCl, plus the test substance. Aspirin, 10 mm and 20 mm, increased net gain of Na+ into the pouch and increased net loss of H+ from the pouch. PD decreased from control of −43.8 my to −31.2 my in response to 20 mm aspirin. Phenylbutazone (200 mg per 100 ml), indomethacin (250 mg per 100 ml), and cortisone acetate (2.5 mg per 100 ml) did not change net movements of electrolytes or alter PD. The net movements of Na+ and H+ increased and PD decreased with the aliphatic alcohols (C1 to C4 ), and this effect was proportional to the molar concentration of the alcohol and to the carbon chain length; the greater the concentration or number of carbon atoms, the greater the effect. Branched chain alcohols (isopropanol and t-butyl alcohol) had less effect on PD and Na+-H+ movements than their straight chain counterparts. Hypertonic solutions of glucose caused a fall in PD and an increase in net Na+-H+ movements. PD was found to be more sensitive of gastric mucosal damage than measurements of ion movement. These studies indicate that the electrolyte and PD response of antral mucosa to alcohols and to anti-inflammatory drugs is generally similar to that found for fundic mucosa.

Comparative Effects of Aliphatic Alcohols on the Gastric Mucosa

Summary We studied the effects of a series of aliphatic alcohols on the gastric mucosa of Heidenhain pouches in four unanesthetized dogs. Transmucosal potential difference (PD). and the net fluxes of H+, Na+, and Cl–, were determined in the presence of various concentrations of methanol, ethanol, n-propanol, and t-butanol. Each alcohol caused a decrease in the PD and an increase in the net flux of Na+ ion into the lumen of the pouch. On a molar basis, t-butanol = n-propanol > ethanol > methanol in their effects on PD and net Na+ flux. Except for t-butanol there was a linear relationship between the ability of the alcohol to alter PD and the oil-water partition coefficient of the alcohol. Low concentrations of ethanol and n-propanol caused a net flux of H+ ion into the pouch (secretion). Higher concentrations of these alcohols and all concentrations of methanol and t-butanol caused a net loss of H+ ion from the pouch. The net flux of Cl- ion tended to parallel and equal the net flux of cations. We conclude that (a) there is no optimal carbon-chain length between C-1 and C-4 for the “damaging” effects of alcohol on the mucosa, (b) the degree of “damage” is best correlated with the lipid solubility of the damaging agent, and (c) there is good correlation between the PD changes and the change in net Na+ flux but not the change in net H+ flux. The authors thank Dr. Lee Forker for valuable advice and for instructions in the use of the scintillation counter. This work was supported by Grant AM 15886 from the U.S. Public Health Service and Veterans Administration Research Funds. A preliminary report of this study was presented at the Amer.

EFFECT OF ADRENALECTOMY AND GLUCOCORTICOIDS ON THE SECRETION AND ABSORPTION OF HYDROGEN ION

The effect of bilateral adrenalectomy on the secretion and absorption of acid in Heidenhain pouch dogs has been investigated. It was found that bilateral adrenalectomy significantly decreased the acid output to varying doses of histamine and did not alter the relation between hydrogen ion concentration and the volume rate of secretion. The absorption of acid by the gastric mucosa was not changed by adrenalectomy. Glucocorticoids but not mineralocorticoids restored secretion to normal levels. These experiments indicate that the impaired secretion of acid in response to histamine in adrenalectomized dogs is not attributable to increased leakage of secreted acid from lumen to blood.

Gastric Emptying in the Cat in Response to Hypertonic Solutions and Tryptophan

Studies of gastric emptying were made in ten cats and four dogs, each with a gastric fistula. Test meals of 50 ml (cats) and 300 ml (dogs) containing phenol red as a nonabsorbable marker were instilled into the stomach via the gastric fistula and were drained by gravity 25–30 min later. The test meals were water, hypertonic solutions (250–1000 mosmol/kg) of casein hydrolysate, D glucose, mannitol, glycine,l-alanine,d-alanine, β-alanine;l- andd-tryptophan andl-phenylalanine in concentrations (osmolalities) ranging from 5 to 100 mM. In the cats, gastric emptying was slowed with increasing osmolality of the test solution, and glucose and mannitol were equipotent but much less effective than glycine or casein hydrolysate. This difference in potencies was not found for these substances in the dogs. Gastric emptying was inhibited by tryptophan in the cats, was stereospecific, and only thel form was effective. This inhibition was not due to the effect of tryptophan on acid secretion. Stereospecificity was not found for the osmoreceptor. Combiningl-tryptophan with another neutral amino acid (l-alanine) inhibited the effect of tryptophan. Phenylalanine in concentrations up to 80 mM had no effect on gastric emptying in the cats. These studies indicate differences between cats and dogs in response to osmotic agents and also show that, like the dog, the cat has a tryptophan receptor.