Many of the drug groups discussed elsewhere in this book have important applications in the treatment of diseases of the gastrointestinal tract and other organs. Other groups are used almost exclusively for their effects on the gut; these are discussed in the following text according to their therapeutic uses.
Drugs Used in Acid-Peptic Diseases
Acid-peptic diseases include gastroesophageal reflux, peptic ulcer (gastric and duodenal), and stress-related mucosal injury. In all these conditions, mucosal erosions or ulceration arise when the caustic effects of aggressive factors (acid, pepsin, bile) overwhelm the defensive factors of the gastrointestinal mucosa (mucus and bicarbonate secretion, prostaglandins, blood flow, and the processes of restitution and regeneration after cellular injury). Over 90% of peptic ulcers are caused by infection with the bacterium Helicobacter pylori or by use of nonsteroidal anti-inflammatory drugs (NSAIDs). Drugs used in the treatment of acid-peptic disorders may be divided into two classes: agents that reduce intragastric acidity and agents that promote mucosal defense.
Agents that Reduce Intragastric Acidity
Physiology of Acid Secretion
The parietal cell contains receptors for gastrin (CCK-B), histamine (H2), and acetylcholine (muscarinic, M3) (Figure 62-1). When acetylcholine (from vagal postganglionic nerves) or gastrin (released from antral G cells into the blood) bind to the parietal cell receptors, they cause an increase in cytosolic calcium, which in turn stimulates protein kinases that stimulate acid secretion from a H+,K+ ATPase (the proton pump) on the canalicular surface.
In close proximity to the parietal cells are gut endocrine cells called enterochromaffin-like (ECL) cells. ECL cells also have receptors for gastrin and acetylcholine, which stimulate histamine release. Histamine binds to the H2 receptor on the parietal cell, resulting in activation of adenylyl cyclase, which increases intracellular cyclic adenosine monophosphate (cAMP) and activates protein kinases that stimulate acid secretion by the H+,K+ ATPase. In humans, it is believed that the major effect of gastrin upon acid secretion is mediated indirectly through the release of histamine from ECL cells rather than through direct parietal cell stimulation. In contrast, acetylcholine provides potent direct parietal cell stimulation.
Antacids have been used for centuries in the treatment of patients with dyspepsia and acid-peptic disorders. They were the mainstay of treatment for acid-peptic disorders until the advent of H2-receptor antagonists and proton pump inhibitors. They continue to be used commonly by patients as nonprescription remedies for the treatment of intermittent heartburn and dyspepsia.
Antacids are weak bases that react with gastric hydrochloric acid to form a salt and water. Their principal mechanism of action is reduction of intragastric acidity. After a meal, approximately 45 mEq/h of hydrochloric acid is secreted. A single dose of 156 mEq of antacid given 1 hour after a meal effectively neutralizes gastric acid for up to 2 hours. However, the acid-neutralization capacity among different proprietary formulations of antacids is highly variable, depending on their rate of dissolution (tablet versus liquid), water solubility, rate of reaction with acid, and rate of gastric emptying.
Sodium bicarbonate (eg, baking soda, Alka Seltzer) reacts rapidly with hydrochloric acid (HCL) to produce carbon dioxide and sodium chloride. Formation of carbon dioxide results in gastric distention and belching. Unreacted alkali is readily absorbed, potentially causing metabolic alkalosis when given in high doses or to patients with renal insufficiency. Sodium chloride absorption may exacerbate fluid retention in patients with heart failure, hypertension, and renal insufficiency. Calcium carbonate (eg, Tums, Os-Cal) is less soluble and reacts more slowly than sodium bicarbonate with HCl to form carbon dioxide and calcium chloride (CaCl2). Like sodium bicarbonate, calcium carbonate may cause belching or metabolic alkalosis. Calcium carbonate is used for a number of other indications apart from its antacid properties (see Chapter 42). Excessive doses of either sodium bicarbonate or calcium carbonate with calcium-containing dairy products can lead to hypercalcemia, renal insufficiency, and metabolic alkalosis (milk-alkali syndrome).
Formulations containing magnesium hydroxide or aluminum hydroxide react slowly with HCl to form magnesium chloride or aluminum chloride and water. Because no gas is generated, belching does not occur. Metabolic alkalosis is also uncommon because of the efficiency of the neutralization reaction. Because unabsorbed magnesium salts may cause an osmotic diarrhea and aluminum salts may cause constipation, these agents are commonly administered together in proprietary formulations (eg, Gelusil, Maalox, Mylanta) to minimize the impact on bowel function. Both magnesium and aluminum are absorbed and excreted by the kidneys. Hence, patients with renal insufficiency should not take these agents long-term.
All antacids may affect the absorption of other medications by binding the drug (reducing its absorption) or by increasing intragastric pH so that the drug’s dissolution or solubility (especially weakly basic or acidic drugs) is altered. Therefore, antacids should not be given within 2 hours of doses of tetracyclines, fluoroquinolones, itraconazole, and iron.
From their introduction in the 1970s until the early 1990s, H2-receptor antagonists (commonly referred to as H2 blockers) were the most commonly prescribed drugs in the world (see Clinical Uses). With the recognition of the role of H pylori in ulcer disease (which may be treated with appropriate antibacterial therapy) and the advent of proton pump inhibitors, the use of prescription H2 blockers has declined markedly.
Chemistry & Pharmacokinetics
Four H2 antagonists are in clinical use: cimetidine, ranitidine, famotidine, and nizatidine. All four agents are rapidly absorbed from the intestine. Cimetidine, ranitidine, and famotidine undergo first-pass hepatic metabolism resulting in a bioavailability of approximately 50%. Nizatidine has little first-pass metabolism. The serum half-lives of the four agents range from 1.1 to 4 hours; however, duration of action depends on the dose given (Table 62-1). H2 antagonists are cleared by a combination of hepatic metabolism, glomerular filtration, and renal tubular secretion. Dose reduction is required in patients with moderate to severe renal (and possibly severe hepatic) insufficiency. In the elderly, there is a decline of up to 50% in drug clearance as well as a significant reduction in volume of distribution.
|Table 62-1 Clinical Comparisons of H2-Receptor Blockers.
BID, twice daily; HS, bedtime.
H2-receptor antagonists continue to be prescribed but proton pump inhibitors (see below) are steadily replacing H2 antagonists for most clinical indications. However, the over-the-counter preparations are heavily used by the public.
Gastroesophageal Reflux Disease (GERD)
Patients with infrequent heartburn or dyspepsia (fewer than 3 times per week) may take either antacids or intermittent H2 antagonists. Because antacids provide rapid acid neutralization, they afford faster symptom relief than H2 antagonists. However, the effect of antacids is short-lived (1-2 hours) compared with H2 antagonists (6-10 hours). H2 antagonists may be taken prophylactically before meals in an effort to reduce the likelihood of heartburn. Frequent heartburn is better treated with twice-daily H2 antagonists (Table 62-1) or proton pump inhibitors. In patients with erosive esophagitis (approximately 50% of patients with GERD), H2 antagonists afford healing in less than 50% of patients; hence proton pump inhibitors are preferred because of their superior acid inhibition.
Peptic Ulcer Disease
Proton pump inhibitors have largely replaced H2 antagonists in the treatment of acute peptic ulcer disease. Nevertheless, H2 antagonists are still sometimes used. Nocturnal acid suppression by H2 antagonists affords effective ulcer healing in most patients with uncomplicated gastric and duodenal ulcers. Hence, all the agents may be administered once daily at bedtime, resulting in ulcer healing rates of more than 80-90% after 6-8 weeks of therapy. For patients with ulcers caused by aspirin or other NSAIDs, the NSAID should be discontinued. If the NSAID must be continued for clinical reasons despite active ulceration, a proton pump inhibitor should be given instead of an H2 antagonist to more reliably promote ulcer healing. For patients with acute peptic ulcers caused by H pylori, H2 antagonists no longer play a significant therapeutic role. H pylori should be treated with a 10- to 14-day course of therapy including a proton pump inhibitor and two antibiotics (see below). This regimen achieves ulcer healing and eradication of the infection in more than 90% of patients. For the minority of patients in whom H pylori cannot be successfully eradicated, H2 antagonists may be given daily at bedtime in half of the usual ulcer therapeutic dose to prevent ulcer recurrence (eg, ranitidine, 150 mg; famotidine, 20 mg).
H2 antagonists are commonly used as over-the-counter agents and prescription agents for treatment of intermittent dyspepsia not caused by peptic ulcer. However, benefit compared with placebo has never been convincingly demonstrated.
Prevention of Bleeding from Stress-Related Gastritis
Clinically important bleeding from upper gastrointestinal erosions or ulcers occurs in 1-5% of critically ill patients as a result of impaired mucosal defense mechanisms caused by poor perfusion. Although most critically ill patients have normal or decreased acid secretion, numerous studies have shown that agents that increase intragastric pH (H2 antagonists or proton pump inhibitors) reduce the incidence of clinically significant bleeding. However, the optimal agent is uncertain at this time. For patients without a nasoenteric tube or with significant ileus, intravenous H2 antagonists are preferable over intravenous proton pump inhibitors because of their proven efficacy and lower cost. Continuous infusions of H2 antagonists are generally preferred to bolus infusions because they achieve more consistent, sustained elevation of intragastric pH.
H2 antagonists are extremely safe drugs. Adverse effects occur in less than 3% of patients and include diarrhea, headache, fatigue, myalgias, and constipation. Some studies suggest that intravenous H2 antagonists (or proton pump inhibitors) may increase the risk of nosocomial pneumonia in critically ill patients.
Mental status changes (confusion, hallucinations, agitation) may occur with administration of intravenous H2 antagonists, especially in patients in the intensive care unit who are elderly or who have renal or hepatic dysfunction. These events may be more common with cimetidine. Mental status changes rarely occur in ambulatory patients.
Cimetidine inhibits binding of dihydrotestosterone to androgen receptors, inhibits metabolism of estradiol, and increases serum prolactin levels. When used long-term or in high doses, it may cause gynecomastia or impotence in men and galactorrhea in women. These effects are specific to cimetidine and do not occur with the other H2 antagonists.
Although there are no known harmful effects on the fetus, H2 antagonists cross the placenta. Therefore, they should not be administered to pregnant women unless absolutely necessary. The H2 antagonists are secreted into breast milk and may therefore affect nursing infants.
H2 antagonists may rarely cause blood dyscrasias. Blockade of cardiac H2 receptors may cause bradycardia, but this is rarely of clinical significance. Rapid intravenous infusion may cause bradycardia and hypotension through blockade of cardiac H2 receptors; therefore, intravenous injections should be given over 30 minutes. H2 antagonists rarely cause reversible abnormalities in liver chemistry.
Cimetidine interferes with several important hepatic cytochrome P450 drug metabolism pathways, including those catalyzed by CYP1A2, CYP2C9, CYP2D6, and CYP3A4 (see Chapter 4). Hence, the half-lives of drugs metabolized by these pathways may be prolonged. Ranitidine binds 4-10 times less avidly than cimetidine to cytochrome P450. Negligible interaction occurs with nizatidine and famotidine.
H2 antagonists compete with creatinine and certain drugs (eg, procainamide) for renal tubular secretion. All of these agents except famotidine inhibit gastric first-pass metabolism of ethanol, especially in women. Although the importance of this is debated, increased bioavailability of ethanol could lead to increased blood ethanol levels.
Proton Pump Inhibitors
Since their introduction in the late 1980s, these efficacious acid inhibitory agents have assumed the major role for the treatment of acid-peptic disorders. Proton pump inhibitors (PPIs) are now among the most widely prescribed drugs worldwide due to their outstanding efficacy and safety.
Chemistry & Pharmacokinetics
Five proton pump inhibitors are available for clinical use: omeprazole, lansoprazole, rabeprazole, pantoprazole, and esomeprazole. All are substituted benzimidazoles that resemble H2 antagonists in structure (Figure 62-3) but have a completely different mechanism of action. Omeprazole is a racemic mixture of R- and S-isomers. Esomeprazole is the S-isomer of omeprazole. All are available in oral formulations. Esomeprazole and pantoprazole are also available in intravenous formulations
|Table 62-2 Pharmacokinetics of Proton Pump InhibItors.
Proton pump inhibitors are administered as inactive prodrugs. To protect the acid-labile prodrug from rapid destruction within the gastric lumen, oral products are formulated for delayed release as acid-resistant, enteric-coated capsules or tablets. After passing through the stomach into the alkaline intestinal lumen, the enteric coatings dissolve and the prodrug is absorbed. For children or patients with dysphagia or enteral feeding tubes, capsules may be opened and the microgranules mixed with apple or orange juice or mixed with soft foods (eg, applesauce). Lansoprazole is also available as a tablet formulation that disintegrates in the mouth, or it may be mixed with water and administered via oral syringe or enteral tube. Omeprazole is also available as a powder formulation (capsule or packet) that contains sodium bicarbonate (1100-1680 mg NaHCO3 ; 304-460 mg of sodium) to protect the naked (non-enteric-coated) drug from acid degradation. When administered on an empty stomach by mouth or enteral tube, this “immediate-release” suspension results in rapid omeprazole absorption (Tmax < 30 minutes) and onset of acid inhibition.
The proton pump inhibitors are lipophilic weak bases (pKa 4-5) and after intestinal absorption diffuse readily across lipid membranes into acidified compartments (eg, the parietal cell canaliculus). The prodrug rapidly becomes protonated within the canaliculus and is concentrated more than 1000-fold by Henderson-Hasselbalch trapping (see Chapter 1). There, it rapidly undergoes a molecular conversion to the active form, a reactive thiophilic sulfenamide cation, which forms a covalent disulfide bond with the H+,K+ ATPase, irreversibly inactivating the enzyme.
From a pharmacokinetic perspective, proton pump inhibitors are ideal drugs: they have a short serum half-life, they are concentrated and activated near their site of action, and they have a long duration of action.
In contrast to H2 antagonists, proton pump inhibitors inhibit both fasting and meal-stimulated secretion because they block the final common pathway of acid secretion, the proton pump. In standard doses, proton pump inhibitors inhibit 90-98% of 24-hour acid secretion (Figure 62-2). When administered at equivalent doses, the different agents show little difference in clinical efficacy. In a crossover study of patients receiving long-term therapy with all five proton pump inhibitors, the mean 24-hour intragastric pH varied from 3.3 (pantoprazole, 40 mg) to 4.0 (esomeprazole, 40 mg) and the mean number of hours the pH was higher than 4 varied from 10.1 (pantoprazole, 40 mg) to 14.0 (esomeprazole, 40 mg).
Gastroesophageal Reflux Disease (GERD)
Proton pump inhibitors are the most effective agents for the treatment of nonerosive and erosive reflux disease, esophageal complications of reflux disease (peptic stricture or Barrett’s esophagus), and extraesophageal manifestations of reflux disease. Once-daily dosing provides effective symptom relief and tissue healing in 85-90% of patients; up to 15% of patients require twice-daily dosing.
GERD symptoms recur in over 80% of patients within 6 months after discontinuation of a proton pump inhibitor. For patients with erosive esophagitis or esophageal complications, long-term daily maintenance therapy with a full-dose or half-dose proton pump inhibitor is usually needed. Many patients with nonerosive GERD may be treated successfully with intermittent courses of proton pump inhibitors or H2 antagonists taken as needed (“on demand”) for recurrent symptoms.
In current clinical practice, many patients with symptomatic GERD are treated empirically with medications without prior endoscopy, ie, without knowledge of whether the patient has erosive or nonerosive reflux disease. Empiric treatment with proton pump inhibitors provides sustained symptomatic relief in 70-80% of patients, compared with 50-60% with H2 antagonists. Because of recent cost reductions, proton pump inhibitors are being used increasingly as first-line therapy for patients with symptomatic GERD.
Sustained acid suppression with twice-daily proton pump inhibitors for at least 3 months is used to treat extraesophageal complications of reflux disease (asthma, chronic cough, laryngitis, and noncardiac chest pain).
Peptic Ulcer Disease
Compared with H2 antagonists, proton pump inhibitors afford more rapid symptom relief and faster ulcer healing for duodenal ulcers and, to a lesser extent, gastric ulcers. All the pump inhibitors heal more than 90% of duodenal ulcers within 4 weeks and a similar percentage of gastric ulcers within 6-8 weeks.
H pylori-Associated Ulcers
For H pylori-associated ulcers, there are two therapeutic goals: to heal the ulcer and to eradicate the organism. The most effective regimens for H pylori eradication are combinations of two antibiotics and a proton pump inhibitor. Proton pump inhibitors promote eradication of H pylori through several mechanisms: direct antimicrobial properties (minor) and—by raising intragastric pH—lowering the minimal inhibitory concentrations of antibiotics against H pylori. The best treatment regimen consists of a 14-day regimen of “triple therapy”: a proton pump inhibitor twice daily; clarithromycin, 500 mg twice daily; and either amoxicillin, 1 g twice daily, or metronidazole, 500 mg twice daily. After completion of triple therapy, the proton pump inhibitor should be continued once daily for a total of 4-6 weeks to ensure complete ulcer healing. Recently, 10 days of “sequential treatment” consisting on days 1-5 of a proton pump inhibitor twice daily plus amoxicillin, 1 g twice daily, and followed on days 6-10 by five additional days of a proton pump inhibitor twice daily, plus clarithromycin, 500 mg twice daily, and tinidazole, 500 mg twice daily, has been shown to be a highly effective treatment regimen.
For patients with ulcers caused by aspirin or other NSAIDs, either H2 antagonists or proton pump inhibitors provide rapid ulcer healing so long as the NSAID is discontinued; however continued use of the NSAID impairs ulcer healing. In patients with NSAID-induced ulcers who require continued NSAID therapy, treatment with a once- or twice-daily proton pump inhibitor more reliably promotes ulcer healing.
Asymptomatic peptic ulceration develops in 10-20% of people taking frequent NSAIDs, and ulcer-related complications (bleeding, perforation) develop in 1-2% of persons per year. Proton pump inhibitors taken once daily are effective in reducing the incidence of ulcers and ulcer complications in patients taking aspirin or other NSAIDs.
Prevention of Rebleeding from Peptic Ulcers
In patients with acute gastrointestinal bleeding due to peptic ulcers, the risk of rebleeding from ulcers that have a visible vessel or adherent clot is increased. Rebleeding of this subset of high-risk ulcers is reduced significantly with proton pump inhibitors administered for 3-5 days either as high-dose oral therapy (eg, omeprazole, 40 mg orally twice daily) or as a continuous intravenous infusion. It is believed that an intragastric pH higher than 6 may enhance coagulation and platelet aggregation. The optimal dose of intravenous proton pump inhibitor needed to achieve and maintain this level of near-complete acid inhibition is unknown; however, initial bolus administration (80 mg) followed by constant infusion (8 mg/h) is commonly recommended.
Proton pump inhibitors have modest efficacy for treatment of nonulcer dyspepsia, benefiting 10-20% more patients than placebo. Despite their use for this indication, superiority to H2 antagonists (or even placebo) has not been conclusively demonstrated.
Prevention of Stress-Related Mucosal Bleeding
As discussed previously (see H2-Receptor Antagonists) proton pump inhibitors (given orally, by nasogastric tube, or by intravenous infusions) may be administered to reduce the risk of clinically significant stress-related mucosal bleeding in critically ill patients. The only proton pump inhibitor approved by the Food and Drug Administration (FDA) for this indication is an oral immediate-release omeprazole formulation, which is administered by nasogastric tube twice daily on the first day, then once daily. For patients with nasoenteric tubes, immediate-release omeprazole may be preferred to intravenous H2 antagonists or proton pump inhibitors because of comparable efficacy, lower cost, and ease of administration.
For patients without a nasoenteric tube or with significant ileus, intravenous H2 antagonists are preferred to intravenous proton pump inhibitors because of their proven efficacy and lower cost. Although proton pump inhibitors are increasingly used, there are no controlled trials demonstrating efficacy or optimal dosing.
Gastrinoma and Other Hypersecretory Conditions
Patients with isolated gastrinomas are best treated with surgical resection. In patients with metastatic or unresectable gastrinomas, massive acid hypersecretion results in peptic ulceration, erosive esophagitis, and malabsorption. Previously, these patients required vagotomy and extraordinarily high doses of H2 antagonists, which still resulted in suboptimal acid suppression. With proton pump inhibitors, excellent acid suppression can be achieved in all patients. Dosage is titrated to reduce basal acid output to less than 5-10 mEq/h. Typical doses of omeprazole are 60-120 mg/d.
Proton pump inhibitors are extremely safe. Diarrhea, headache, and abdominal pain are reported in 1-5% of patients, although the frequency of these events is only slightly increased compared with placebo. Proton pump inhibitors do not have teratogenicity in animal models; however, safety during pregnancy has not been established.
Acid is important in releasing vitamin B12 from food. A minor reduction in oral cyanocobalamin absorption occurs during proton pump inhibition, potentially leading to subnormal B12 levels with prolonged therapy. Acid also promotes absorption of food-bound minerals (iron, calcium, zinc); however, no mineral deficiencies have been reported with proton pump inhibitor therapy. Recent case-control studies have suggested a modest increase in the risk of hip fracture in patients taking proton pump inhibitors over a long term compared with matched controls. Although a causal relationship is unproven, proton pump inhibitors may reduce calcium absorption or inhibit osteoclast function. Pending further studies, patients who require long-term proton pump inhibitors—especially those with risk factors for osteoporosis—should have monitoring of bone density and should be provided calcium supplements.
Respiratory and Enteric Infections
Gastric acid is an important barrier to colonization and infection of the stomach and intestine from ingested bacteria. Increases in gastric bacterial concentrations are detected in patients taking proton pump inhibitors, which is of unknown clinical significance. Some studies have reported an increased risk of both community-acquired respiratory infections and nosocomial pneumonia among patients taking proton pump inhibitors.
A small increased risk of enteric infections may exist in patients taking proton pump inhibitors, especially when traveling in underdeveloped countries. Hospitalized patients may have an increased risk for Clostridium difficile infection.
Potential Problems Due to Increased Serum Gastrin
Gastrin levels are regulated by intragastric acidity. Acid suppression alters normal feedback inhibition so that median serum gastrin levels rise 1.5- to 2-fold in patients taking proton pump inhibitors. Although gastrin levels remain within normal limits in most patients, they exceed 500 pg/mL (normal, < 100 pg/mL) in 3%. Upon stopping the drug, the levels normalize within 4 weeks.
The rise in serum gastrin levels in patients receiving long-term therapy with proton pump inhibitors raises a theoretical concern because gastrin may stimulate hyperplasia of ECL cells. In female rats given proton pump inhibitors for prolonged periods, gastric carcinoid tumors developed in areas of ECL hyperplasia. Although humans who take proton pump inhibitors for a long time may exhibit ECL hyperplasia in response to hypergastrinemia, carcinoid tumor formation has not been documented. At present, routine monitoring of serum gastrin levels is not recommended in patients receiving prolonged proton pump inhibitor therapy.
Other Potential Problems Due to Decreased Gastric Acidity
Among patients infected with H pylori, long-term acid suppression leads to increased chronic inflammation in the gastric body and decreased inflammation in the antrum. Concerns have been raised that increased gastric inflammation may accelerate gastric gland atrophy (atrophic gastritis) and intestinal metaplasia—known risk factors for gastric adenocarcinoma. A special FDA Gastrointestinal Advisory Committee concluded that there is no evidence that prolonged proton pump inhibitor therapy produces the kind of atrophic gastritis (multifocal atrophic gastritis) or intestinal metaplasia that is associated with increased risk of adenocarcinoma. Routine testing for H pylori is not recommended in patients who require long-term proton pump inhibitor therapy. Long-term proton pump inhibitor therapy is associated with the development of small benign gastric fundic-gland polyps in a small number of patients, which may disappear after stopping the drug and are of uncertain clinical significance.
Decreased gastric acidity may alter absorption of drugs for which intragastric acidity affects drug bioavailability, eg, ketoconazole, itraconazole, digoxin, and atazanavir. All proton pump inhibitors are metabolized by hepatic P450 cytochromes, including CYP2C19 and CYP3A4. Because of the short half-lives of proton pump inhibitors, clinically significant drug interactions are rare. Omeprazole may inhibit the metabolism of warfarin, diazepam, and phenytoin. Esomeprazole also may decrease metabolism of diazepam. Lansoprazole may enhance clearance of theophylline. Rabeprazole and pantoprazole have no significant drug interactions.
Mucosal Protective Agents
The gastroduodenal mucosa has evolved a number of defense mechanisms to protect itself against the noxious effects of acid and pepsin. Both mucus and epithelial cell-cell tight junctions restrict back diffusion of acid and pepsin. Epithelial bicarbonate secretion establishes a pH gradient within the mucous layer in which the pH ranges from 7 at the mucosal surface to 1-2 in the gastric lumen. Blood flow carries bicarbonate and vital nutrients to surface cells. Areas of injured epithelium are quickly repaired by restitution, a process in which migration of cells from gland neck cells seals small erosions to rees