Chapter 49 Stomach
The stomach begins as a dilation in the tubular embryonic foregut during the fifth week of gestation. By the seventh week, it descends, rotates, and further dilates with a disproportionate elongation of the greater curvature into its normal anatomic shape and position. Following birth, it is the most proximal abdominal organ of the alimentary tract. The most proximal region of the stomach is called the cardia, which attaches to the esophagus. Immediately proximal to the cardia is a physiologically competent lower esophageal sphincter. Distally, the pylorus connects the distal stomach (antrum) to the proximal duodenum. Although the stomach is fixed at the gastroesophageal (GE) junction and pylorus, its large midportion is mobile. The fundus represents the superiormost part of the stomach and is floppy and distensible. The stomach is bounded superiorly by the diaphragm and laterally by the spleen. The body of the stomach represents the largest portion and is also referred to as the corpus. The body also contains most of the parietal cells and is bounded on the right by the relatively straight lesser curvature and on the left by the longer greater curvature. At the angularis incisura, the lesser curvature abruptly angles to the right. It is here that the body of the stomach ends and the antrum begins. Another important anatomic angle (angle of His) is that formed by the fundus with the left margin of the esophagus (Fig. 49-1).
FIGURE 49-1 Divisions of the stomach.
(From Yeo C: Shackelford’s surgery of the alimentary tract, ed 6, Philadelphia, 2007, WB Saunders.)
Most of the stomach resides within the upper abdomen. The left lateral segment of the liver covers a large portion of the stomach anteriorly. The diaphragm, chest, and abdominal wall bound the remainder of the stomach. Inferiorly, the stomach is attached to the transverse colon, spleen, caudate lobe of the liver, diaphragmatic crura, and retroperitoneal nerves and vessels. Superiorly, the GE junction is found about 2 to 3 cm below the diaphragmatic esophageal hiatus in the horizontal plane of the seventh chondrosternal articulation, a plane only slightly cephalad to that containing the pylorus. The gastrosplenic ligament attaches the proximal greater curvature to the spleen.
The celiac artery provides most of the blood supply to the stomach (Fig. 49-2). There are four main arteries—the left and right gastric arteries along the lesser curvature and the left and right gastroepiploic arteries along the greater curvature. In addition, a substantial quantity of blood may be supplied to the proximal stomach by the inferior phrenic arteries and by the short gastric arteries from the spleen. The largest artery to the stomach is the left gastric artery, and it is not uncommon (15% to 20%) for an aberrant left hepatic artery to originate from it. Consequently, proximal ligation of the left gastric artery occasionally results in acute left-sided hepatic ischemia. The right gastric artery arises from the hepatic artery (or the gastroduodenal artery). The left gastroepiploic artery originates from the splenic artery and the right gastroepiploic originates from the gastroduodenal artery. The extensive anastomotic connection between these major vessels ensures that in most cases, the stomach will survive if three out of four arteries are ligated, provided that the arcades along the greater and lesser curvatures are not disturbed. In general, the veins of the stomach parallel the arteries. The left gastric (coronary) and right gastric veins usually drain into the portal vein. The right gastroepiploic vein drains into the superior mesenteric vein and the left gastroepiploic vein drains into the splenic vein.
The lymphatic drainage of the stomach parallels the vasculature and drains into four zones of lymph nodes (Fig. 49-3). The superior gastric group drains lymph from the upper lesser curvature into the left gastric and paracardial nodes. The suprapyloric group of nodes drains the antral segment on the lesser curvature of the stomach into the right suprapancreatic nodes. The pancreaticolienal group of nodes drains lymph high on the greater curvature into the left gastroepiploic and splenic nodes. The inferior gastric and subpyloric group of nodes drains lymph along the right gastroepiploic vascular pedicle. All four zones of lymph nodes drain into the celiac group and into the thoracic duct. Although these lymph nodes drain different areas of the stomach, gastric cancers may metastasize to any of the four nodal groups, regardless of the cancer location. In addition, the extensive submucosal plexus of lymphatics accounts for the fact that there is frequently microscopic evidence of malignant cells several centimeters from gross disease.
As shown in Figure 49-4, the extrinsic innervation of the stomach is parasympathetic (via the vagus) and sympathetic (via the celiac plexus). The vagus nerve originates in the vagal nucleus in the floor of the fourth ventricle and traverses the neck in the carotid sheath to enter the mediastinum, where it divides into several branches around the esophagus. These branches coalesce above the esophageal hiatus to form the left and right vagus nerves. It is not uncommon to find more than two vagal trunks at the distal esophagus. At the GE junction, the left vagus is anterior, and the right vagus is posterior (LARP).
FIGURE 49-4 Vagal innervation of the stomach. The line of division for truncal vagotomy is shown; it is above the hepatic and celiac branches of the left and right vagus nerves, respectively. The line of division for selective vagotomy is shown; this is below the hepatic and celiac branches.
(From Mercer D, Liu T: Open truncal vagotomy. Oper Tech Gen Surg 5:8–85, 2003.)
The left vagus gives off the hepatic branch to the liver and then continues along the lesser curvature as the anterior nerve of Latarjet. Although not shown, the so-called criminal nerve of Grassi is the first branch of the right or posterior vagus nerve; it is recognized as a potential cause of recurrent ulcers when left undivided. The right nerve gives a branch off to the celiac plexus and then continues posteriorly along the lesser curvature. A truncal vagotomy is performed above the celiac and hepatic branches of the vagi, whereas a selective vagotomy is performed below. A highly selective vagotomy is performed by dividing the crow’s feet to the proximal stomach while preserving the innervation of the antral and pyloric parts of the stomach. Most (90%) of the vagal fibers are afferent, carrying stimuli from the gut to the brain. Efferent vagal fibers originate in the dorsal nucleus of the medulla and synapse with neurons in the myenteric and submucosal plexuses. These neurons use acetylcholine as their neurotransmitter and influence gastric motor function and gastric secretion. In contrast, the sympathetic nerve supply comes from T5 to T10, traveling in the splanchnic nerve to the celiac ganglion. Postganglionic fibers then travel with the arterial system to innervate the stomach.
The intrinsic or enteric nervous system of the stomach consists of neurons in Auerbach’s and Meissner’s autonomic plexuses. In these locations, cholinergic, serotoninergic, and peptidergic neurons are present. However, the function of these neurons remains poorly understood. Nevertheless, a number of neuropeptides have been localized to these neurons; these include acetylcholine, serotonin, substance P, calcitonin gene–related peptide (CGRP), bombesin, cholecystokinin (CCK), and somatostatin. Consequently, it is an oversimplification to think of the stomach as only containing parasympathetic (cholinergic input) and sympathetic (adrenergic input) supply. Moreover, the parasympathetic nervous system contains adrenergic neurons and the sympathetic system also contains cholinergic neurons.
The stomach is covered by peritoneum, which forms the outer serosa of the stomach. Below it is the thicker muscularis propria, or muscularis externa, which is made up of three layers of smooth muscles. The middle layer of smooth muscle is circular and is the only complete muscle layer of the stomach wall. At the pylorus, this middle circular muscle layer becomes progressively thicker and functions as a true anatomic sphincter. The outer muscle layer is longitudinal and continuous with the outer layer of longitudinal esophageal smooth muscle. Within the layers of the muscularis externa is a rich plexus of autonomic nerves and ganglia, called Auerbach’s myenteric plexus. The submucosa lies between the muscularis externa and mucosa and is a collagen-rich layer of connective tissue that is the strongest layer of the gastric wall. In addition, it contains the rich anastomotic network of blood vessels and lymphatics and Meissner’s plexus of autonomic nerves. The mucosa consists of surface epithelium, lamina propria, and muscularis mucosae. The latter is on the luminal side of the submucosa and is probably responsible for the rugae that greatly increase epithelial surface area. It also marks the microscopic boundary for invasive and noninvasive gastric carcinoma. The lamina propria represents a small connective tissue layer and contains capillaries, vessels, lymphatics, and nerves necessary to support the surface epithelium.
Gastric mucosa consists of columnar glandular epithelia. The cellular populations (and functions) of the cells forming this glandular epithelium vary based on their location in the stomach (Table 49-1). The glandular epithelium is divided into cells that secrete products into the gastric lumen for digestion (parietal cells, chief cells, mucus-secreting cells) and cells that control function (gastrin-secreting G cells, somatostatin-secreting D cells) cells. In the cardia, the mucosa is arranged in branched glands and the pits are short. In the fundus and body, the glands are more tubular and the pits are longer. In the antrum, the glands are more branched. The luminal ends of the gastric glands and pits are lined with mucus-secreting surface epithelial cells, which extend down into the necks of the glands for variable distances. In the cardia, the glands are predominantly mucus-secreting. In the body, the glands are mostly lined from the neck to the base with parietal and chief cells (Fig. 49-5). There are a few parietal cells in the fundus and proximal antrum, but none in the cardia or prepyloric antrum. The endocrine G cells are present in greatest quantity in the antral glands.
|Parietal||Body||Secretion of acid and intrinsic factor|
|Surface epithelial||Diffuse||Mucus, bicarbonate, prostaglandins (?)|
|Gastric mucosal interneurons||Body, antrum||Gastrin-releasing peptide|
|Enteric neurons||Diffuse||Calcitonin gene–related peptide, others|
The principal function of the stomach is to prepare ingested food for digestion and absorption as it is propulsed into the small intestine. The initial period of digestion requires that solid components of a meal be stored for several hours while they undergo a reduction in size and break down into their basic metabolic constituents.
Receptive relaxation of the proximal stomach enables the stomach to function as a storage organ. Receptive relaxation refers to the process whereby the proximal portion of the stomach relaxes in anticipation of food intake. This relaxation enables liquids to pass easily from the stomach along the lesser curvature, whereas the solid food settles along the greater curvature of the fundus. In contrast to liquids, emptying of solid food is facilitated by the antrum, which pumps solid food components into and through the pylorus. The antrum and pylorus function in a coordinated fashion, allowing entry of food components into the duodenum and also returning material to the proximal stomach until it is suitable for delivery into the duodenum.
In addition to storing food, the stomach begins digestion of a meal. Starches undergo enzymatic breakdown through the activity of salivary amylase. Peptic digestion metabolizes a meal into fats, proteins, and carbohydrates by breaking down cell walls. Although the duodenum and proximal small intestine are primarily responsible for digestion of a meal, the stomach facilitates this process.
Gastric function is under neural (sympathetic and parasympathetic) and hormonal control (peptides or amines that interact with target cells in the stomach). An understanding of the roles of endocrine and neural regulation of digestion is critical to understanding gastric physiology. Abnormal secretion of gastrin and pepsin was thought to be the major causative factor in peptic ulcer disease (PUD). The discovery of Helicobacter pylori (H. pylori) and the effect of this organism on ulcer disease has rendered moot many of the theoretical rationales for acid hypersecretion. A general understanding of gastric physiology and the specific impact of peptides on acid secretion, however, is still critical to understanding the physiologic effects of gastric surgical procedures on digestion. We will initially focus here on peptide regulation of gastric function and then describe the interactions of these peptides with neural inputs in regard to acid secretion and gastric function.
Gastrin is produced by G cells located in the gastric antrum (see Table 49-1). It is synthesized as a prepropeptide and undergoes post-translational processing to produce biologically reactive gastrin peptides. Several molecular forms of gastrin exist. G-34 (big gastrin), G-17 (little gastrin), and G-14 (minigastrin) have been identified. However, 90% of antral gastrin is released as the 17–amino acid peptide, although G-34 predominates in the circulation because its metabolic half-life is longer than that of G-17. The pentapeptide sequence contained at the carboxyl terminus of gastrin is the biologically active component and is identical to that found on another gut peptide, CCK. CCK and gastrin differ by tyrosine sulfation sites. The release of gastrin is stimulated by food components in a meal, especially protein digestion products. Luminal acid inhibits the release of gastrin. In the antral location, somatostatin and gastrin release are functionally linked, and an inverse reciprocal relationship exists between these two peptides.1
Gastrin is the major hormonal regulator of the gastric phase of acid secretion following a meal. Histamine, released from enterochromaffin-like (ECL) cells, is also a potent stimulant of acid release from the parietal cell. Gastrin also has considerable trophic effects on the parietal cells and gastric ECL cells. Prolonged hypergastrinemia from any cause leads to mucosal hyperplasia and an increase in the number of ECL cells and, under some circumstances, is associated with the development of gastric carcinoid tumors.2
The detection of hypergastrinemia may suggest a pathologic state of acid hypersecretion but generally is the result of treatment with agents to lower acid secretion, such as proton pump inhibitors. Table 49-2 lists common causes of chronic hypergastrinemia. Hypergastrinemia that results from the administration of acid-lowering drugs is an appropriate response caused by loss of feedback inhibition of gastrin release by luminal acid. Lack of acid causes a reduction in somatostatin release, which in turn causes increased release of gastrin from antral G cells. Hypergastrinemia can also occur in the setting of pernicious anemia or uremia, or following surgical procedures such as vagotomy or retained gastric antrum after gastrectomy. In contrast, gastrin levels increase inappropriately in patients with gastrinoma (Zollinger-Ellison syndrome [ZES]). These gastrin-secreting tumors are not located in the antrum and secrete gastrin autonomously.
|ULCEROGENIC CAUSES||NONULCEROGENIC CAUSES|
|Antral G-cell hyperplasia or hyperfunction||Antisecretory agents (PPIs)|
|Retained excluded antrum||Atrophic gastritis|
|Zollinger-Ellison syndrome||Pernicious anemia|
|Gastric outlet obstruction||Acid-reducing procedure (vagotomy)|
|Short-gut syndrome||Helicobacter pylori infection|
|Chronic renal failure|
Gastrin initiates its biologic actions by activation of surface membrane receptors. These receptors are members of the classic G protein–coupled seven–transmembrane-spanning receptor family and are classified as type A or B CCK receptors. The gastrin or CCK-B receptor has high affinity for gastrin and CCK, whereas the type A CCK receptors have an affinity for sulfated CCK analogues and a low affinity for gastrin. Binding of gastrin with the CCK-B receptor has been associated with elevated intracellular calcium levels.
Somatostatin is produced by D cells and exists endogenously as the 14– or 28–amino acid peptide. The predominant molecular form in the stomach is somatostatin-14. It is produced by diffuse neuroendocrine cells located in the fundus and antrum. In these locations, D cell cytoplasmic extensions have direct contact with the parietal cells and G cells, where it presumably exerts its actions through paracrine effects on acid secretion and gastrin release.3 Somatostatin is able to inhibit parietal cell acid secretion directly but can also indirectly inhibit acid secretion through inhibition of gastrin release and downregulation of histamine release from ECL cells. The principal stimulus for somatostatin release is antral acidification, whereas acetylcholine from vagal fibers inhibits its release.
Somatostatin receptors are also seven–transmembrane-spanning receptors. Binding of somatostatin with its receptors is coupled to one or more inhibitory guanine nucleotide–binding proteins. Parietal cell somatostatin receptors appear to be a single subunit of glycoproteins with a molecular weight of 99 kDa, with equal affinity for somatostatin-14 and somatostatin-28. Somatostatin can inhibit parietal cell secretion through G protein–dependent and G protein–independent mechanisms. However, the ability of somatostatin to exert its inhibitory actions on cellular function is primarily thought to be mediated through the inhibition of adenylate cyclase, with a resultant reduction in cyclic AMP levels.
Bombesin was discovered in 1970 in an extract prepared from skin of the amphibian Bombina bombina (European fire-bellied toad). Its mammalian counterpart is gastrin-releasing peptide (GRP). GRP is particularly prominent in nerves ending in the acid-secreting and gastrin-secreting portions of the stomach and is found in the circular muscular layer. In the antral mucosa, GRP stimulates gastrin and somatostatin release by binding to receptors located on the G and D cells, respectively. It is rapidly cleared from the circulation by a neutral endopeptidase and has a half-life of approximately 1.4 minutes. Peripheral administration of exogenous GRP stimulates gastric acid secretion, whereas central administration in the ventricles inhibits acid secretion. The inhibitory pathway activated is not mediated by a humoral factor, is unaffected by vagotomy, and appears to involve the sympathetic nervous system.
Histamine plays a prominent role in parietal cell stimulation. Administration of histamine 2 (H2) receptor antagonists almost completely abolishes gastric acid secretion in response to gastrin and acetylcholine. This suggests that histamine may be a necessary intermediary of gastrin- and acetylcholine-stimulated acid secretion. Histamine is stored in the acidic granules of ECL cells and in resident mast cells. Its release is stimulated by gastrin, acetylcholine, and epinephrine following receptor-ligand interactions on ECL cells. In contrast, somatostatin inhibits gastrin-stimulated histamine release through interactions with somatostatin receptors located on the ECL cell. Thus, the ECL cell plays an essential role in parietal cell activation that possesses stimulatory and inhibitory feedback pathways that modulate the release of histamine and therefore acid secretion.
Ghrelin is a 28–amino acid peptide predominantly produced by endocrine cells of the oxyntic mucosa of the stomach, with substantially lower amounts from the bowel, pancreas, and other organs. Removal of the acid-producing part of the stomach decreases circulating ghrelin by 80%. Ghrelin appears to be under endocrine and metabolic control, has a diurnal rhythm, likely plays a major role in the neuroendocrine and metabolic responses to changes in nutritional status, and may be a major anabolic hormone.
In human volunteers, ghrelin administration enhances appetite and increases food intake. In patients who have undergone a gastric bypass, ghrelin levels are 77% lower than those of matched obese controls, a finding not seen after other forms of antiobesity surgery. Although the mechanism responsible for suppression of ghrelin levels after gastric bypass is unknown, this suggests that ghrelin may be responsive to the normal flow of nutrients across the stomach. Other studies have suggested that ghrelin leads to a switch toward glycolysis and away from fatty acid oxidation, which would favor fat deposition. It appears that ghrelin is upregulated in times of negative energy balance and downregulated in times of positive energy balance, although the precise role of ghrelin in energy metabolism remains unclear. Ghrelin may come to have a role in the treatment and prevention of obesity.
Gastric acid secretion by the parietal cell is regulated by three local stimuli—acetylcholine, gastrin, and histamine. These three stimuli account for basal and stimulated gastric acid secretion. Acetylcholine is the principal neurotransmitter modulating acid secretion and is released from the vagus and parasympathetic ganglion cells. Vagal fibers innervate not only parietal cells but also G cells and ECL cells to modulate release of their peptides. Gastrin has hormonal effects on the parietal cell and stimulates histamine release. Histamine has paracrine-like effects on the parietal cell and, as shown in Figure 49-6, plays a central role in the regulation of acid secretion by the parietal cell after its release from ECL cells. As depicted, somatostatin exerts inhibitory actions on gastric acid secretion. Release of somatostatin from antral D cells is stimulated in the presence of intraluminal acid to a pH of 3 or lower. After its release, somatostatin inhibits gastrin release through paracrine effects and also modifies histamine release from ECL cells. In some patients with PUD, this negative feedback response is defective. Consequently, the precise state of acid secretion by the parietal cell is dependent on the overall influence of the positive and negative stimuli.
FIGURE 49-6 Central role of the ECL cell in regulation of acid secretion by the parietal cell. As shown, ingestion of a meal stimulates vagal fibers to release acetylcholine (cephalic phase). Binding of acetylcholine to M3 receptors located on the ECL cell, parietal cell, and G cell results in the release of histamine, hydrochloric acid, and gastrin, respectively. Binding of acetylcholine to M3 receptors on D cells results in the inhibition of somatostatin release. Following a meal, G cells are also stimulated to release gastrin, which interacts with receptors located on ECL cells and parietal cells to cause the release of histamine and hydrochloric acid (gastric phase). Release of somatostatin from D cells decreases histamine release and gastrin release from ECL cells and G cells, respectively. In addition, somatostatin inhibits parietal cell acid secretion (not shown). The principal stimulus for the activation of D cells is antral luminal acidification (not shown).
(From Yeo C: Shackelford’s surgery of the alimentary tract, ed 6, Philadelphia, 2007, WB Saunders.)
In the absence of food, there is always a basal level of acid secretion that is approximately 10% of maximal acid output. Under basal conditions, 1 to 5 mmol/hr of hydrochloric acid is secreted, and this is reduced after vagotomy or H2 receptor blockade. Thus, it appears likely that basal acid secretion is caused by a combination of cholinergic and histaminergic input.
Ingestion of food is the physiologic stimulus for acid secretion. Three phases of the acid secretory response to a meal have been described, cephalic, gastric, and intestinal. These three phases are interrelated and occur concurrently, not consecutively.
The cephalic phase originates with the sight, smell, thought, and/or taste of food, which stimulates neural centers in the cortex and hypothalamus. Although the exact mechanisms whereby senses stimulate acid secretion remain to be fully elucidated, it is hypothesized that several sites are stimulated in the brain. These higher centers transmit signals to the stomach by the vagus nerves, which release acetylcholine that in turn activates muscarinic receptors located on target cells. Acetylcholine directly increases acid secretion by the parietal cells and can inhibit and stimulate gastrin release, the net effect being a slight increase in gastrin levels. Although the intensity of the acid secretory response in the cephalic phase surpasses that of the other phases, it accounts for only 20% to 30% of the total volume of gastric acid produced in response to a meal because of the short duration of the cephalic phase.
The gastric phase of acid secretion begins when food enters the gastric lumen. Digestion products of ingested food interact with microvilli of antral G cells to stimulate gastrin release. Food also stimulates acid secretion by causing mechanical distention of the stomach. Gastric distention activates stretch receptors in the stomach to elicit the long vagovagal reflex arc. It is abolished by proximal gastric vagotomy and is, at least in part, independent of changes in serum gastrin levels. However, antral distention also causes gastrin release in humans, a reflex that has been called the pyloro-oxyntic reflex. In humans, mechanical distention of the stomach accounts for about 30% to 40% of the maximal acid secretory response to a peptone meal, with the remainder caused by gastrin release. The entire gastric phase accounts for most (60% to 70%) of meal-stimulated acid output because it lasts until the stomach is empty.
The intestinal phase of gastric secretion remains poorly understood but appears to be initiated by entry of chyme into the small intestine. It occurs after gastric emptying and lasts as long as partially digested food components remain in the proximal small bowel. It accounts for only 10% of the acid secretory response to a meal and does not appear to be mediated by serum gastrin levels. It has been hypothesized that a distinct acid stimulatory peptide hormone (entero-oxyntin) released from small bowel mucosa may mediate the intestinal phase of acid secretion.
The two second messengers principally involved in stimulation of acid secretion by parietal cells are intracellular cyclic AMP (cAMP) and calcium. Synthesis of these two messengers in turn activates protein kinases and phosphorylation cascades. The intracellular events following ligand binding to receptors on the parietal cell are shown in Figure 49-7. Histamine causes an increase in intracellular cAMP, which activates protein kinases to initiate a cascade of phosphorylation events that culminate in activation of H+, K+-ATPase. In contrast, acetylcholine and gastrin stimulate phospholipase C, which converts membrane-bound phospholipids into inositol triphosphate (IP3) to mobilize calcium from intracellular stores. Increased intracellular calcium activates other protein kinases that ultimately activate H+, K+-ATPase in a similar fashion to initiate the secretion of hydrochloric acid.
FIGURE 49-7 Intracellular signaling events in a parietal cell. As shown, histamine binds to H2 receptors, stimulating adenylate cyclase through a G protein–linked mechanism. Adenylate cyclase activation causes an increase in intracellular cAMP levels, which in turn activates protein kinases. Activated protein kinases stimulate a phosphorylation cascade, with a resultant increase in levels of phosphoproteins that activate the proton pump. Activation of the proton pump leads to extrusion of cytosolic hydrogen in exchange for extracytoplasmic potassium. In addition, chloride is secreted through a chloride channel located on the luminal side of the membrane. Gastrin binds to type B cholecystokinin receptors and acetylcholine binds to M3 receptors. Following the interaction of gastrin or acetylcholine with their receptors, phospholipase C is stimulated through a G protein–linked mechanism to convert membrane-bound phospholipids into inositol triphosphate (IP3). IP3 stimulates the release of calcium from intracellular calcium stores, leading to an increase in intracellular calcium that in turn activates protein kinases, which activate the H+,K+-ATPase. ATP, Adenosine triphosphate; ATPase, adenosine triphosphatase; cAMP, cyclic adenosine monophosphate; Gi, inhibitory guanine nucleotide protein; Gs,stimulatory guanine nucleotide protein; PIP2, phosphatidylinositol 4,5- diphosphate; PLC, phospholipase C.
(From Yeo C: Shackelford’s surgery of the alimentary tract, ed 6, Philadelphia, 2007, WB Saunders.)
H+, K+-ATPase is the final common pathway for gastric acid secretion by the parietal cell. It is composed of two subunits, a catalytic α-subunit (100 kDa) and a glycoprotein β-subunit (60 kDa). During the resting, or nonsecreting, state, gastric parietal cells store H+, K+-ATPase within intracellular tubulovesicular elements. Cellular relocation of the proton pump subunits through cytoskeletal rearrangements must occur in order for acid secretion to increase in response to stimulatory factors. The subsequent insertion and heterodimer assembly of the H+, K+-ATPase subunits into the microvilli of the secretory canaliculus causes an increase in gastric acid secretion. A KCl efflux pathway must exist to supply potassium to the extracytoplasmic side of the pump. Cytosolic hydrogen is secreted by H+, K+-ATPase in exchange for extracytoplasmic potassium (see Fig. 49-7), which is an electroneutral exchange and therefore does not contribute to the transmembrane potential difference across the parietal cell. Secretion of chloride is accomplished through a chloride channel moving chloride from the parietal cell cytoplasm to the gastric lumen. The secretion or exchange of hydrogen for potassium, however, does require energy in the form of adenosine triphosphate (ATP) because hydrogen is being secreted against a gradient of more than a million-fold. Because of this large energy requirement, the parietal cell also has the largest mitochondrial content of any mammalian cell, with a mitochondrial compartment representing 34% of its cell volume. In response to a secretagogue, the parietal cell undergoes a conformational change, and a several-fold increase in the canalicular surface area occurs (Fig. 49-8). In contrast to stimulated acid secretion, cessation of acid secretion requires endocytosis of H+, K+-ATPase, with regeneration of cytoplasmic tubulovesicles containing the subunits, and this occurs through a tyrosine-based signal. The tyrosine-containing sequence is located on the cytoplasmic tail of the β-subunit and is highly homologous to the motif responsible for internalization of the transferrin receptor.
FIGURE 49-8 Diagrammatic representation of resting and stimulated parietal cells. Note the morphologic transformation between the nonsecreting parietal cell and stimulated parietal cell, with increases in secretory canalicular membrane surface area.
More than 1 billion parietal cells are found in the normal human stomach and are responsible for secreting about 20 mmol/hr of hydrochloric acid in response to a protein meal. Each individual parietal cell secretes 3.3 billion hydrogen ions/second, and there is a linear relationship between maximal acid output and parietal cell number. However, gastric acid secretory rates may be altered in patients with upper gastrointestinal (GI) disease. For example, gastric acid is often increased in patients with duodenal ulcer or gastrinoma, whereas it is decreased in patients with pernicious anemia, gastric atrophy, gastric ulcer, or gastric cancer. The lower secretory rates observed in gastric ulcer patients are typically for proximal gastric ulcers, whereas distal, antral, or prepyloric ulcers are associated with acid secretory rates similar to those in duodenal ulcer patients.
Gastric acid thus plays a critical role in the digestion of a meal. It is required to convert pepsinogen into pepsin, elicits the release of secretin from the duodenum, and limits colonization of the upper GI tract with bacteria.
The diversity of mechanisms that stimulate acid secretion has resulted in the development of many site-specific drugs aimed at decreasing acid output by the parietal cell. The best-known site-specific antagonists are the group collectively known as the H2 receptor antagonists. The most potent of the H2 receptor antagonists is famotidine, followed by ranitidine, nizatidine, and cimetidine. The half-life for famotidine is 3 hours and approximately 1.5 hours for the others. All undergo hepatic metabolism, are excreted by the kidney, and do not differ much in bioavailability.
The newest class of antisecretory agents is the proton pump inhibitors (PPIs). These substituted benzimidazoles, of which omeprazole is a prime example, inhibit acid secretion more completely because of their irreversible inhibition of the proton pump. These PPIs are weak acids with a pKa of 4 and therefore become selectively localized in the secretory canaliculus of the parietal cell, which is the only structure in the body with a pH lower than 4. After oral administration, these agents are absorbed into the bloodstream as prodrugs and then selectively concentrate in the secretory canaliculus. At low pH, they become ionized and activated, with the formation of an active sulfur group. Because the proton pump is located on the luminal surface, the transmembrane pump proteins are also exposed to acid or low pH. The cysteine residues on the α-subunit form a covalent disulfate bond with activated benzimidazoles, which irreversibly inhibits the proton pump. Because of the covalent nature of this bond, these PPIs have more prolonged inhibition of gastric acid secretion than H2 blockers. In order for recovery of acid secretion to occur, new protein pumps need to be synthesized. As a result, these agents have a longer duration of action than their plasma half-life, with the intragastric pH being maintained higher than 3 for 18 hours or longer.
One notable side effect of all antisecretory agents is the elevation of serum gastrin levels. Serum gastrin levels are higher after treatment with PPIs than with H2 receptor antagonists. This effect is accompanied by hyperplasia of G cells and ECL cells when these agents are administered chronically. Chronic administration of omeprazole has been found to cause ECL hyperplasia that could progress to carcinoid tumors in rats.3 This effect, however, was not specific for omeprazole and was reproduced by other agents that caused prolonged inhibition of acid secretion and resultant hypergastrinemia.
Gastric juice is the result of secretion by the parietal cells, chief cells, and mucus cells, in addition to swallowed saliva and duodenal refluxate. The electrolyte composition of parietal and nonparietal gastric secretion varies with the rate of gastric secretion. Parietal cells secrete an electrolyte solution that is isotonic with plasma and contains 160 mmol/liter. The pH of this solution is 0.8. The lowest intraluminal pH commonly measured in the stomach is 2 because of dilution of the parietal cell secretion by other gastric secretions, which also contain sodium, potassium, and bicarbonate.
Intrinsic factor is a 60-kDa mucoprotein secreted by the parietal cell that is essential for the absorption of vitamin B12 in the terminal ileum. It is secreted in amounts that far exceed those necessary for vitamin B12 absorption. In general, its secretion parallels that of gastric acid secretion, yet the secretory response is not necessarily linked to acid secretion. For example, PPIs do not block intrinsic factor secretion in humans nor do they alter the absorption of labeled vitamin B12. Intrinsic factor deficiency can develop in the setting of pernicious anemia or in patients undergoing total gastrectomy, and both groups of patients require vitamin B12 supplementation.
Pepsinogens are proteolytic proenzymes with a molecular weight of 42,500 that are secreted by the glands of the gastroduodenal mucosa. Two types of pepsinogens are secreted. Group 1 pepsinogens are secreted by chief cells and by mucus neck cells located in the glands of the acid-secreting portion of the stomach. Group 2 pepsinogens are produced by surface epithelial cells throughout the acid-secreting portion of the stomach, antrum, and proximal duodenum. Consequently, group 1 pepsinogens are secreted by the same glands that secrete acid, whereas group 2 pepsinogens are secreted by acid-secreting and gastrin-secreting mucosa. In the presence of acid, both forms of pepsinogen are converted to pepsin by removal of a short amino-terminal peptide. Pepsins become inactivated at a pH higher than 5, although group 2 pepsinogens are active over a wider range of pH values than group 1 pepsinogens. As a result, group 2 pepsinogens may be involved in peptic digestion in the presence of increased gastric pH, which commonly occurs in the setting of stress or in patients with gastric ulcer.
Mucus and bicarbonate combine to neutralize gastric acid at the gastric mucosal surface. They are secreted by the surface mucus cells and mucus neck cells located in the acid-secreting and antral portions of the stomach. Mucus is a viscoelastic gel that contains approximately 85% water and 15% glycoproteins. It provides a mechanical barrier to injury by contributing to the unstirred layer of water found at the luminal surface of the gastric mucosa. It also acts as an impediment to ion movement from the lumen to the apical cell membrane and is relatively impermeable to pepsins. Mucus is in a constant state of flux because it is secreted continuously by mucosal cells on the one hand and solubilized by luminal pepsin on the other. Mucus production is stimulated by vagal stimulation, cholinergic agonists, prostaglandins, and some bacterial toxins. In contrast, anticholinergic drugs and nonsteroidal anti-inflammatory drugs (NSAIDs) inhibit its secretion. H. pylori, on the other hand, secretes various proteases and lipases that break down mucin and impair the protective function of the mucus layer.
In the acid-secreting portion of the stomach, bicarbonate secretion is an active process, whereas in the antrum, active and passive secretion of bicarbonate occur. The magnitude of bicarbonate secretion, however, is considerably less than acid secretion. Although the luminal pH is 2, the pH observed at the surface epithelial cell is usually 7. The pH gradient found at the epithelial surface is a result of the unstirred layer of water in the mucus gel and of the continuous secretion of bicarbonate by the surface epithelial cells. Gastric cell surface pH remains higher than 5 until the luminal pH is less than 1.4. However, the luminal pH in duodenal ulcer patients is frequently less than 1.4, so the cell surface is exposed to a lower pH in these patients. This reduction in pH may reflect a reduction in gastric bicarbonate secretion and decreased duodenal bicarbonate secretion, and may explain why some duodenal ulcer patients have a higher relapse rate after treatment.
Gastric motility is regulated by extrinsic and intrinsic neural mechanisms and by myogenic control. The extrinsic neural controls are mediated through parasympathetic (vagus) and sympathetic pathways, whereas the intrinsic controls involve the enteric nervous system (see earlier, “Anatomy”). In contrast, myogenic control resides in the excitatory membranes of the gastric smooth muscle cells.
The electrical basis of gastric motility begins with the depolarization of pacemaker cells located in the midbody of the stomach, along the greater curvature. Once initiated, slow waves travel at 3 cycles/minute in a circumferential and antegrade fashion toward the pylorus. In addition to these slow waves, gastric smooth muscle cells are capable of producing action potentials, which are associated with larger changes in membrane potential than slow waves. In comparison to slow waves, which are not associated with gastric contractions, action potentials are associated with actual muscle contractions. During fasting, the stomach goes through a cyclical pattern of electrical activity composed of slow waves and electrical spikes, which has been termed the myoelectric migrating complex (MMC). Each cycle of the MMC lasts 90 to 120 minutes. The net effects of the MMC are frequent clearance of gastric contents during periods of fasting. The exact regulatory mechanisms of MMC activities are unknown, but these activities remain intact after vagal denervation.
Ingestion of a meal results in a decrease in the resting tone of the proximal stomach and fundus, referred to as receptive relaxation and gastric accommodation, respectively. Because these reflexes are mediated by the vagus nerves, interruption of vagal innervation to the proximal stomach, such as by truncal vagotomy or proximal gastric vagotomy, can eliminate these reflexes, with resultant early satiety and rapid emptying of ingested liquids. In addition to its storage function, the stomach is responsible for the mixing and grinding of ingested solid food particles. This activity involves repetitive forceful contractions of the mid and antral portions of the stomach, causing food particles to be propelled against a closed pylorus, with subsequent retropulsion of solids and liquids. The net effect is a thorough mixing of solids and liquids and sequential shearing of solid food particles to smaller than 1 mm.
The emptying of gastric contents is under the influence of well-coordinated neural and hormonal mediators. Systemic factors, such as anxiety, fear, depression, and exercise, can affect the rate of gastric motility and emptying. Additionally, the chemical and mechanical properties and temperature of the intraluminal contents can influence the rate of gastric emptying. In general, liquids empty more rapidly than solids and carbohydrates empty more readily than fats. An increase in the concentration or acidity of liquid meals causes a delay in gastric emptying. In addition, hot and cold liquids tend to empty at a slower rate than ambient temperature fluids. These responses to luminal stimuli are regulated by the enteric nervous system. Osmoreceptors and pH-sensitive receptors in the proximal small bowel have also been shown to be involved in the activation of feedback inhibition of gastric emptying. Inhibitory peptides proposed to be active in this setting include CCK, glucagon, vasoactive intestinal peptide, and gastric inhibitory polypeptide.
Symptoms of abnormal gastric motility are nausea, fullness, early satiety, abdominal pain, and discomfort. Although mechanical obstruction can and should be ruled out with upper endoscopy or radiographic contrast studies, objective evaluation of a patient with a suspected motility disorder can be accomplished with gamma scintigraphy, real-time ultrasound, and magnetic resonance imaging (MRI). Gastric motility disorders usually encountered in clinical practice are gastric dysmotility following vagotomy, delayed gastric emptying associated with diabetes mellitus, and gastric motility dysfunction related to H. pylori infection. Vagotomy results in loss of receptive relaxation and gastric accommodation in response to meal ingestion, with resultant early satiety, postprandial bloating, accelerated emptying of liquids, and delay in emptying of solids. Clinical manifestations of diabetic gastropathy, which can occur in insulin-dependent or insulin-independent patients, closely resemble the clinical picture of postvagotomy gastroparesis. Furthermore, structural changes have been identified in the vagus nerve of patients with diabetes, suggesting that a diabetic autonomic neuropathy may be responsible. However, the metabolic effects of diabetes have also been implicated. Specifically, hyperglycemia has been shown to cause a decrease in contractility of the gastric antrum, increase in pyloric contractility, and suppression of the migrating motor complex (MMC). Suppression of MMC activity is thought to be responsible for the accumulation of gastric bezoars seen in some diabetic patients. In contrast, hyperinsulinemia, which is often associated with non–insulin-dependent diabetes, may play a role in the gastroparesis seen in non–insulin-dependent diabetes because it also leads to suppression of MMC activity.4
H. pylori–infected patients with nonulcer dyspepsia have also been demonstrated to have impaired gastric emptying accompanied by a reduction in gastric compliance.5 In rats, lipopolysaccharide derived from H. pylori causes a reduction in gastric emptying of a liquid meal for up to 12 hours by an unknown mechanism. Regardless of the cause of gastroparesis, treatment consists of prokinetic agents, such as metoclopramide and erythromycin, which have been shown to have some benefit, although the evidence is more compelling in diabetics.
There are a number of ways to assess gastric emptying. The saline load test is perhaps the simplest and is accomplished by instilling a known volume of saline into the stomach and aspirating the amount remaining at a certain time. Alternatively, fluoroscopic procedures can also provide information on gastric emptying and may reveal mechanical causes that could contribute to a delay, such as gastric outlet obstruction. However, computerized radionucleotide scans are more commonly used to assess gastric emptying. This can be done with radiolabeled liquids or with a radiolabeled solid meal. After a mechanical obstruction has been ruled out, gastric emptying studies using these radionucleotide scans can be particularly helpful for patients with gastric atony from an associated illness such as diabetes or in postgastrectomy patients.
Gastric barrier function depends on physiologic and anatomic factors. Blood flow plays a critical role in gastric mucosal defense by providing nutrients and delivering oxygen to ensure that the intracellular processes that underlie mucosal resistance to injury can proceed unabated. Decreased gastric mucosal blood flow has minimal effects on lesion production until it approaches 50% of normal. When blood flow is reduced by more than 75%, marked mucosal injury results, which is exacerbated in the presence of luminal acid. After damage occurs, injured surface epithelial cells are replaced rapidly by the migration of surface mucus cells located along the basement membranes. This process is referred to as restitution or reconstitution. It occurs within minutes and does not require cell division.
Exposure of the stomach to noxious agents causes a reduction in the potential difference across the gastric mucosa. In normal gastric mucosa, the potential difference across the mucosa is −30 to −50 mV and results from the active transport of chloride into the lumen and sodium into the blood by the activity of Na+,K+-ATPase. Damage disrupts the tight junctions between mucosal cells, causing the epithelium to become leaky to ions (i.e., Na+ and Cl−) and a resultant loss of the high transepithelial electrical resistance normally found in gastric mucosa. In addition, damaging agents such as NSAIDs or aspirin possess carboxyl groups that are nonionized at a low intragastric pH because they are weak acids. Consequently, they readily enter the cell membranes of gastric mucosal cells because they are now lipid-soluble, whereas they will not penetrate the cell membranes at neutral pH because they are ionized. On entry into the neutral pH environment found in the cytosol, they become reionized, will not exit the cell membrane, and are toxic to the mucosal cells.
The estimated prevalence of PUD ranges from 5% to 15% in Western populations, with a lifetime incidence of almost 10%.6 Although the incidence and hospitalization rate for PUD have been decreasing since the 1980s, it remains one of the most prevalent and costly GI diseases. Medical costs associated with PUD are an estimated $5.65 billion annually. An estimated 15,000 operations are performed each year on patients hospitalized with PUD. Significant progress has been made over the past 2 decades, with total admissions for PUD decreasing by almost 30%. Admissions for complications of ulcer disease have also been decreasing, which has led to a significant decrease in ulcer-related mortality, from 3.9% in 1993 to 2.7% in 2006.7 Although overall mortality remains low, this still represents over 4000 deaths caused by PUD each year.
The role of surgery in the treatment of ulcer disease has also decreased, primarily caused by a marked decline in elective surgical therapy for chronic disease because the percentage of patients who require emergent surgery for complicated disease has remained constant, at 7% of hospitalized patients.7 This represents over 11,000 surgical procedures annually.
Much of this decline in ulcer incidence and the need for hospitalization has stemmed from increased knowledge of ulcer pathogenesis. Specifically, the role of H. pylori has been defined and the risks of chronic NSAID use have been better elucidated. An increase in H. pylori eradication will hopefully result in a decrease of not just elective surgical procedures, but also a decline in complications and mortality from emergent complications.
Peptic ulcers are caused by increased aggressive factors, decreased defensive factors, or both.8 This in turn leads to mucosal damage and subsequent ulceration. Protective (or defensive) factors include mucosal bicarbonate secretion, mucus production, blood flow, growth factors, cell renewal, and endogenous prostaglandins. Damaging (or aggressive) factors include hydrochloric acid secretion, pepsins, ethanol ingestion, smoking, duodenal reflux of bile, ischemia, NSAIDs, hypoxia and, most notably, H. pylori infection.
It is now believed that 90% of duodenal ulcers and approximately 75% of gastric ulcers are associated with H. pylori infection. When this organism is eradicated as part of ulcer treatment, ulcer recurrence is extremely rare. H. pylori is a spiral or helical gram-negative rod with four to six flagella that resides in gastric-type epithelium within or beneath the mucus layer. This location protects the bacteria from acid and antibiotics. Its shape and flagella aid its movement through the mucus layer and it produces enzymes that help it adapt to this hostile environment. Most notably, it is a potent producer of urease, which is capable of splitting urea into ammonia and bicarbonate, creating an alkaline microenvironment in the setting of an acidic gastric milieu. The secretion of this enzyme, however, facilitates detection of the organism. H. pylori is microaerophilic and can only live in gastric epithelium. Thus, it can also be found in heterotopic gastric mucosa in the proximal esophagus, Barrett’s esophagus, gastric metaplasia in the duodenum, within a Meckel’s diverticulum, and heterotopic gastric mucosa in the rectum.
Peptic ulcers are also strongly associated with antral gastritis. Studies done before the H. pylori era have demonstrated that almost all peptic ulcer patients have histologic evidence of antral gastritis. It was later found that the only patients with gastric ulcers and no gastritis were those ingesting aspirin. It is now recognized that most cases of histologic gastritis are caused by H. pylori infection. Even 25% of patients with an NSAID-associated ulcer have evidence of a histologic antral gastritis, as opposed to 95% of those with non–NSAID-associated ulcers. In most cases, the infection tends to be confined initially to the antrum and results in antral inflammation. Other evidence supporting a causal role for H. pylori in histologic gastritis comes from two separate volunteer physicians who ingested inocula of H. pylori after first confirming normal gross and microscopic gastric mucosa. Both developed gastric H. pylori infection. Acute inflammation was observed histologically on days 5 and 10. By 2 weeks, it had been replaced by chronic inflammation with evidence of a mononuclear cell infiltration. These two reports provide documentation that H. pylori can cause histologic gastritis. However, histologic gastritis does not necessarily equate with symptoms of dyspepsia.
H. pylori infection usually occurs in childhood, and spontaneous remission is rare. There is an inverse relationship between infection and socioeconomic status. The reasons for this remain poorly understood but seem to be the result of factors such as sanitary conditions, familial clustering, and crowding. This likely explains why developing countries have a comparatively higher rate of H. pylori infection, especially in children.
A number of studies have demonstrated what appears to be a steady linear increase in the acquisition of H. pylori infection with age, especially in the United States and northern European nations. In the United States, H. pylori prevalence also varies among racial and ethnic groups.
H. pylori infection is associated with a number of common upper GI disorders, but most infected individuals are asymptomatic. Normal U.S. blood donors have an overall prevalence of about 20% to 55%. H. pylori infection is almost always present in the setting of active chronic gastritis and is present in most duodenal (>90%) and gastric (60% to 90%) ulcer patients. Noninfected gastric ulcer patients tend to be NSAID users. There is weaker association with nonulcer dyspepsia. In addition, most gastric cancer patients have current or past H. pylori infection. Although the association between H. pylori and cancer is strong, no causal relationship has been proven. H. pylori–induced chronic gastritis and intestinal metaplasia, however, are thought to play a role. There is also a strong association between +lymphoma and H. pylori infection. Regression of these lymphomas has been demonstrated after eradication of the organism.
Limited data are available to estimate the lifetime risk of PUD in patients with H. pylori infection. In a longitudinal study from Australia with a mean evaluation period of 18 years, 15% of H. pylori–positive subjects developed verified duodenal ulcer as compared with 3% of seronegative individuals. In a 10-year study of asymptomatic gastritis patients, 11% of patients with histologic gastritis developed PUD over a 10-year period, compared with only 1% of those without gastritis. Another factor implicating a causative role for H. pylori and ulcer formation is that eradication of H. pylori dramatically reduces ulcer recurrence. Many prospective trials have shown that patients with H. pylori infection and non-NSAID ulcer disease who have documented eradication of the organism almost never (<2%) develop recurrent ulcers.
Hospitalizations for bleeding upper GI lesions have increased together with the increased use of NSAIDs. The risk for bleeding and ulceration is proportional to the daily dosage of NSAIDs. The risk also increases with age older than 60 years, patients having a prior GI event, or concurrent use of steroids or anticoagulants. Consequently, the ingestion of NSAIDs remains an important factor in ulcer pathogenesis, especially in regard to the development of complications and death. More than 3 million people in the United States use NSAIDs daily. When compared with the general population, NSAID users have a 2- to 10-fold increased risk for GI complications.
The risk for mucosal injury or ulceration is roughly proportional to the anti-inflammatory effect associated with each NSAID. In comparison to H. pylori ulcers, which are more frequently found in the duodenum, NSAID-induced ulcers are more often found in the stomach. H. pylori ulcers are also almost always associated with chronic active gastritis, whereas gastritis is not frequently found with an NSAID-induced ulcer, occurring only about 25% of the time. When NSAID use is discontinued, the ulcers usually do not recur.
Acid plays an important but likely noncausative role in the formation of ulcers. In duodenal ulcers, there is a large overlap of acid levels between ulcer patients and normal subjects. Almost 70% of patients with duodenal ulcers have an acid output within the normal range. Acid levels alone provide little information and, as such, acid secretory testing is of little value in establishing a diagnosis of duodenal ulcer.
For types I and IV gastric ulcers, which are not associated with excessive acid secretion, acid acts as an important cofactor, exacerbating the underlying ulcer damage and attenuating the ability of the stomach to heal. For patients with type II or III gastric ulcers, gastric acid hypersecretion seems to be more common, and consequently they behave more like duodenal ulcers.
Patients suffering from duodenal ulcer disease can present in various ways. The most common symptom associated with duodenal ulcer disease is midepigastric abdominal pain that is usually well localized. The pain is generally tolerable and frequently relieved by food. The pain may be episodic, seasonal in the spring and fall, and worse during periods of emotional stress. Many patients do not seek medical attention until they have had the disease for many years. When the pain becomes constant, this suggests that there is deeper penetration of the ulcer. Referral of pain to the back is usually a sign of penetration into the pancreas. Diffuse peritoneal irritation is usually a sign of free perforation.
History and physical examination are of limited value in distinguishing between gastric and duodenal ulceration. Routine laboratory studies include a complete blood count, liver chemistries, and serum creatinine, serum amylase, and calcium levels. A serum gastrin level should also be obtained in patients with ulcers that are refractory to medical therapy or require surgery. An upright chest radiograph is usually performed when ruling out perforation. The two principal means of diagnosing duodenal ulcers are upper GI radiography and fiberoptic endoscopy. Upper GI is less expensive, and most (90%) ulcers can be diagnosed accurately by this means. However, about 5% of ulcers that appear radiographically benign are malignant. H. pylori testing should also be done in all patients with suspected PUD.
H. pylori can be diagnosed by mucosal biopsy, but noninvasive tests offer an effective screening tool and do not require an endoscopic procedure. Serology is the test of choice for initial diagnosis when endoscopy is not required. However, if endoscopy is to be performed, the rapid urease assay and histology are both excellent options.
There are various enzyme-linked immunosorbent assay (ELISA) laboratory-based tests available and some rapid office-based immunoassays. Serology has a 90% sensitivity and specificity rate. Antibody titers can remain high for 1 year or longer; consequently, this test cannot be used to assess eradication after therapy.
The carbon-labeled urea breath test is based on the ability of H. pylori to hydrolyze urea. Its sensitivity and specificity are both higher than 95%. The urea breath test is less expensive than endoscopy and samples the entire stomach. False-negative results can occur if the test is done too soon after treatment, so it is usually best to perform this test 4 weeks after therapy is finished. The urea breath test is the method of choice to document eradication.
Endoscopy can also be performed with biopsy samples of gastric mucosa, followed by histologic visualization of H. pylori using routine hematoxylin and eosin stains or with special stains (e.g., silver, Giemsa, Genta stains) for improved visibility. Sensitivity is approximately 95% and specificity 99%. This test is widely available and affords the clinician the ability to assess the severity of gastritis and confirm the presence or absence of the organism.
Culturing of gastric mucosa obtained at endoscopy can also be performed to diagnose H. pylori. The sensitivity is approximately 80% and specificity 100%. However, it requires laboratory expertise, is not widely available and is relatively expensive, and diagnosis requires up to 3 to 5 days. Nevertheless, it provides the opportunity to perform antibiotic sensitivity testing on isolates, if needed.
Diagnosis of duodenal ulcer by upper GI radiography requires the demonstration of barium within the ulcer crater, which is usually round or oval and may or may not be surrounded by edema. This study is useful to determine the location and depth of penetration of the ulcer and the extent of deformation from chronic fibrosis. A characteristic barium radiograph of a peptic ulcer is shown in Figure 49-9. The ability to detect ulcers on radiography requires the technical skills and abilities of the radiologist but is also dependent on the size and location of the ulcer. With single-contrast radiographic techniques, as many as 50% of duodenal ulcers may be missed, whereas with double-contrast studies, 80% to 90% of ulcer craters can be detected.
Antiulcer drugs fall into three broad categories—those targeted against H. pylori, those that reduce acid levels by decreasing secretion or chemical neutralization, and those that increase the mucosal protective barrier. In patients with PUD and H. pylori infection, the focus of therapy is on eradication of the bacteria. In addition to medications, lifestyle changes, such as smoking cessation, discontinuing NSAIDs and aspirin, and avoiding coffee and alcohol, all help promote ulcer healing.
Antacids are the oldest form of therapy for PUD. Antacids reduce gastric acidity by reacting with hydrochloric acid, forming a salt and thereby raising the gastric pH. Antacids differ greatly in their buffering ability, absorption, taste, and side effects. Magnesium antacids tend to be the best buffers but can cause significant diarrhea, whereas acids precipitated with phosphorus can occasionally result in hypophosphatemia and sometimes constipation. They are most effective when ingested 1 hour after a meal because they can be retained in the stomach and exert their buffering action for longer periods. If taken on an empty stomach, antacids are emptied rapidly and have only a transient buffering effect. Dosages of 200 to 1000 mmol/day produce minimal side effects and result in approximately 80% ulcer healing at 1 month. Although antacids may heal duodenal ulcers with an efficacy comparable to that observed with H2 receptor antagonists, many patients have found large frequent doses to be unacceptable.
The H2 receptor antagonists are structurally similar to histamine. Variations in ring structure and side chains cause differences in potency and side effects. Currently available H2 receptor antagonists differ in their potency but only modestly in half-life and bioavailability. All undergo hepatic metabolism and are excreted by the kidney. Famotidine is the most potent and cimetidine is the weakest. Continuous IV infusion of H2 receptor antagonists has been shown to produce more uniform acid inhibition than intermittent administration. Many randomized controlled trials have indicated that all H2 receptor antagonists result in duodenal ulcer healing rates from 70% to 80% after 4 weeks and from 80% to 90% after 8 weeks of therapy.
The most potent antisecretory agents are PPIs. These agents negate all types of acid secretion from all types of secretogogues. As a result, they provide a more complete and prolonged inhibition of acid secretion than H2 receptor antagonists. H2 receptor antagonists and PPIs are effective at night, but PPIs are more effective during the day. PPIs have a healing rate of 85% at 4 weeks and 96% at 8 weeks and produce more rapid healing of ulcers than standard H2 receptor antagonists (14% advantage at 2 weeks and 9% advantage at 4 weeks). PPIs require an acidic environment within the gastric lumen to become activated; thus, using antacids or H2 receptor antagonists in combination with PPIs could have deleterious effects by promoting an alkaline environment and thereby preventing activation of the PPIs. Consequently, antacids and H2 receptor antagonists should not be used in combination with PPIs.
Sucralfate is structurally related to heparin but does not have any anticoagulant effects. It has been shown to be effective in the treatment of ulcer disease, although its exact mechanism of action is not entirely understood. It is an aluminum salt of sulfated sucrose that dissociates under the acidic conditions in the stomach. It is hypothesized that the sucrose polymerizes and binds to protein in the ulcer crater to produce a protective coating that can last for up to 6 hours. It has also been suggested that it may bind and concentrate endogenous basic fibroblast growth factor, which appears to be important for mucosal healing. Duodenal ulcer healing after 4 to 6 weeks of treatment with sucralfate is superior to placebo and comparable that of H2 receptor antagonists such as cimetidine.
Prior to the discovery of H. pylori infection as the causative agent in over 95% of duodenal peptic ulcers, the primary form of treatment was the reduction of acid in the stomach, with or without increasing the protective barrier with drugs such as sucralfate. After it became clear that increased acid secretion was an effect of H. pylori infection, there was a paradigm shift that saw PUD as an infectious disease, rather than a consequence of pathologic acid secretion. Accordingly, treatment philosophy has shifted to focus on eradication of the infectious agent.
Current therapy is twofold in its approach, combining antibiotics against H. pylori with antacid medications. The primary goal of the antacids is to promote short-term healing by reducing pathologic acid levels and improve symptoms. H. pylori eradication helps with initial healing, but its primary efficacy is in preventing recurrence. There have been numerous trials comparing eradication therapy with ulcer-healing drugs alone or no treatment. Eradication of H. pylori has shown recurrence rates as low as 2%, with initial healing as high as 90%. This compares with recurrence rates of up to 25% with ulcer-healing medications alone. One review has analyzed the results of these trials and further validated the role of antibiotics in the treatment of H. pylori–positive duodenal ulcers.10 Both eradication therapy and ulcer-healing drugs alone have shown initial healing rates higher than 80%. Eradication therapy has resulted in long-term recurrence of less than 15%, in contrast to 64% recurrence in patients treated with only a short initial course of ulcer-healing drugs. Patients could achieve low recurrence rates similar to those of eradication therapy, but only if they were maintained on their antacid regiment long term, as opposed to a 1- or 2-week course of eradication therapy.
Given all these findings across multiple studies, and according to the recommendations of the American Gastroenterological Association, European Helicobacter pylori Study Group, and National Institutes of Health (NIH), the treatment of H. pylori–positive peptic duodenal ulcer disease is triple therapy aimed at the eradication of H. pylori, along with acid suppression (Box 49-1). This includes an antisecretory agent, now most commonly a PPI, although histamine antagonists are still used, along with two antibiotics, usually amoxicillin with clarithromycin or metronidazole, given for a 2-week course. Side effects, which are generally mild and resolve with cessation of treatment, include diarrhea, nausea and vomiting, rash, and altered taste. For the 10% of patients with refractory disease, quadruple therapy with the addition of bismuth is recommended.
Box 49-1 National Institutes of Health Consensus Panel Recommendations for Helicobacter pylori Treatment
Ulcer disease was once very much the purview of the general surgeon, with ulcer surgery forming a major part of general surgery practice. With the shift in understanding of the disease from one primarily of aberrant acid physiology to one of infectious disease, this has changed significantly, with the overwhelming majority of ulcer patients being treated and cured medically. The surgeon’s role now is primarily to treat the approximately 20% of patients who have a complication from their disease, which includes hemorrhage, perforation, and obstruction (Box 49-2). Frequently included in discussions of complicated ulcer disease is the intractable ulcer. Although intractable disease no doubt exists, its definition is nebulous and exactly when and what type of surgical intervention it requires remain primarily a matter of judgment.
Box 49-2 Surgical Treatment Recommendations for Complications Related to Peptic Duodenal Ulcer Disease
Upper GI bleeding remains a relatively common problem, with an annual incidence of approximately 1/1000.11 Most nonvariceal bleeding (70%) is attributable to peptic ulcers. Most bleeding will stop spontaneously and requires no intervention; persistent bleeding, however, is associated with a 6% to 8% mortality.
The primary clinical criteria that predict persistent bleeding or rebleeding after initial cessation of bleeding, and thus increased mortality, are increased age, decreased hemoglobin (<10 g/dL) at presentation, shock, melena, and need for blood transfusion. Patients who meet any of these criteria should be considered as high risk.11
Almost all patients with an acute upper GI bleed should have endoscopy within 24 hours. Although the data are not conclusive, early endoscopy has been shown to be a cost-effective strategy by triaging patients to more rapid intervention, if warranted, and by identifying low-risk patients without the need for prolonged observation (and therefore earlier hospital discharge).
The initial approach to an upper GI bleed is similar to the approach to a trauma patient. Large-bore IV access, rapid restoration of intravascular volume with fluid and blood products as the clinical situation dictates, and close monitoring for signs of rebleeding are all essential to effective management of these patients. The role of nasogastric (NG) lavage remains an area of debate; however, it can be useful as a predictor of high-risk patients and as an aid for later endoscopic intervention. Patients with bright red blood on NG lavage, as opposed to clear or coffee ground lavage, are at much higher risk for persistent bleeding or rebleeding and warrant endoscopic intervention. Furthermore, the NG tube can be used to lavage the stomach and duodenum prior to endoscopy, removing clot and old blood that could obscure visualization of the source of bleeding. Given its relatively low risk and potential benefit, NG tube placement should be part of the treatment algorithm for these patients once appropriate intravascular access has been established and resuscitation begun.
Patients who are noted to have active bleeding, via an arterial jet or oozing, an adherent clot, or a visible vessel within the ulcer, are at high risk and intervention is required. Patients without active bleeding, no visible vessel, and a clean ulcer base are low risk and do not require further intervention. All patients undergoing endoscopic examination should be tested for their H. pylori status. For the high-risk patient requiring intervention, the best initial approach is endoscopic control, which results in primary hemostasis in approximately 90% of patients. The most common method of control is injection of a vasoconstrictor at the site of bleeding. However, with this method alone, primary hemostasis rates are high but up to 30% of patients have rebleeding. This has led to the development of new techniques, including use of a second vasoconstrictor or sclerosing agent, thermal coagulation, and placement of clips at the site of bleeding. A 2007 meta-analysis has compared the use of epinephrine alone to epinephrine plus any second technique.12 A dual approach, when compared with epinephrine alone, showed better primary hemostasis, reduction in rebleeding rate, lower rate of surgery, and decreased mortality.
Importantly, both thermal coagulation and mechanical clips, in single-method comparisons with epinephrine, have shown significant superiority with respect to primary hemostasis and rebleeding. A meta-analysis comparing either of these two techniques alone to the use of dual methods has shown no significant difference, except in cases of active arterial bleeding, in which case use of a second method was superior. Although the cost and complications from epinephrine remain small, the use of dual methods does have slightly higher, although still less than 1%, complication rates (e.g., necrosis, perforation), than any single technique. Current 2003 guidelines for endoscopic control of bleeding advocate the use of epinephrine plus an additional method.11 As more data become available, it may be demonstrated that thermal or mechanical treatment alone can be used for most patients. For patients who have rebleeding, repeat endoscopy does not increase their mortality and should be attempted prior to surgical intervention.
All high-risk patients should be placed in a monitored setting, preferably an intensive care unit, until all bleeding has stopped for 24 hours. As part of the 2003 consensus guidelines, all high-risk patients should be placed on an IV PPI, with an initial bolus followed by continuous infusion or intermittent dosing for up to 72 hours. When compared with a histamine blocker and placebo, the IV PPI showed lower rebleeding rates, lower rate of emergency surgery, and decreased mortality.13 Patients deemed high risk based on clinical factors who are awaiting endoscopy should probably begin therapy, even prior to endoscopy.
Despite the use of PPIs and improved methods of endoscopic control, 5% to 10% of patients will have persistent bleeding that requires surgical intervention. The vessel most likely to be bleeding is the gastroduodenal artery because of erosion from a posterior ulcer. The duodenum is opened longitudinally, with the incision carried across the pylorus. The vessel is oversewn, with a three-point U stitch technique, which effectively ligates the main vessel along with any smaller branches. One must be careful to avoid incorporating the common bile duct into the stitch. The duodenotomy is closed transversely to avoid narrowing (Fig. 49-10).
FIGURE 49-10 A-E, Heineke-Mikulicz pyloroplasty.
(From Soreide JA, Soreide A: Pyloroplasty. Oper Tech Gen Surg 5:65–72, 2003.)