Gastrointestinal pharmacology: practical tips for the esophagologist

Gastroesophageal reflux disease (GERD) is primarily a motor disorder, and its pathogenesis is multifactorial. As a consequence, treatment should be able to address the underlying pathophysiology. Proton pump inhibitors (PPIs) are the mainstay of medical therapy for GERD, but these drugs only provide the control of symptoms and lesions without curing the disease. However, continuous acid suppression with PPIs is recommended for patients with Barrett's esophagus because of their potential chemopreventive effects. In addition to the antisecretory activity, these compounds display several pharmacological properties, often overlooked in clinical practice. PPIs can indeed affect gastric motility, exert a mucosal protective effect, and an antioxidant, anti‐inflammatory, and antineoplastic activity, also protecting cancer cells from developing chemo‐ or radiotherapeutic resistance. Even in the third millennium, current pharmacologic approaches to address GERD are limited. Reflux inhibitors represent a promise unfulfilled, effective and safe prokinetics are lacking, and antidepressants, despite being effective in selected patients, give rise to adverse events in a large proportion of them. While waiting for new drug classes (like potassium‐competitive acid blockers), reassessing old drugs (namely alginate‐containing formulations), and paving the new avenue of esophageal mucosal protection are, at the present time, the only reliable alternatives to acid suppression.

Gastroesophageal reflux disease (GERD) is primarily a motor disorder, and its pathogenesis is multifactorial. 1 As a consequence, treatment should be able to address the underlying pathophysiology. Unfortunately, this is not the case with proton pump inhibitors (PPIs) that only provide a palliative approach to the disease, targeting the stomach rather than the esophagus, where abnormal acid exposure is not secondary to gastric acid hyper-secretion, which has been documented in only a small subset of GERD patients. 1 Currently, the mainstay of medical therapy for GERD is the administration of PPIs. Their use can actually be diagnostic for GERD in the setting of typical symptoms of heartburn and regurgitation. 2 PPIs work by inhibiting proton pumps, which are activated during a meal. By irreversibly binding and inhibiting the pumps, these drugs significantly decrease gastric acid secretion, raise the intragastric pH, and reduce the symptoms of GERD. 3 Other therapeutic options act to either decrease acid production (histamine doi: 10.1111/nyas.14447 2-receptor antagonists (H 2 RAs)), or buffer and/or block gastric acid from refluxing into the esophagus (alginate-containing formulations) among others. A newer antisecretory drug, vonoprazan (belonging to the class of potassium-competitive acid blockers (P-CABs)), is being used in Japan and other Asian countries, as well as in South America, and is currently being evaluated for use in North America and Europe. [4][5][6] Our review stems from presentations given at a roundtable during the OESO World Conference held in Beijing, China, on November 7-9, 2019. Each topic, selected by the Organizing Committee, was assigned to a given expert, who carried out an independent systematic search of the relevant literature, using Medline/PubMed, Embase, and the Cochrane databases. Search outputs were distilled, paying more attention to systematic reviews and meta-analyses (where available), representing the best evidence. Owing to time constraints, each speaker had to select the key issues of a given topic and this choice is reflected in the present review.

PPIs: are they still the drugs of choice for long-term management of GERD?
Over the years, there has been an increasing number of papers, describing possible risks of PPIs. Some of these risks have been disproven (e.g., the decreased efficacy of clopidogrel), while others persist in the literature as possible adverse outcomes. 2,[7][8][9] The risk of adverse effects of PPIs has led to a reconsideration of their use (especially in the long-term) and reevaluation of their role in the management of GERD. These risks include, among others, chronic kidney disease (CKD), dementia, community-acquired pneumonia, Clostridium difficile, other bacterial gastrointestinal (GI) illnesses, mineral malabsorption, and bone fracture. 8,9 Much of the currently available data regarding risks of PPIs are based on retrospective cohort studies. Together with a large amount of retrospective data ascribing various adverse effects to PPIs, there are several studies showing that there is less risk or none at all. [8][9][10] One of these studies is a recent prospective analysis of patients randomized to either being on a PPI or no PPI. 11 The patients were randomized to receive rivaroxaban with aspirin, rivaroxaban only, or aspirin only. 11 Patients not already on a PPI were randomized to PPI or placebo. 11 The study lasted 3 years and ultimately found that, in this specific study population, there were no increased associated risks other than a category defined as "other enteric infections" (P = 0.04). 11 While one could argue that this study is not generalizable to the entire population on the basis of the study design (the study included patients who had atherosclerosis and who were on antiplatelet agents, and were only evaluated over a 3-year period), these data are the strongest evidence to date to assess the risk associated with PPI use. 11 One significant long-term effect of PPI use has been CKD and warrants a more in-depth appraisal of the data. 12 In 2016, there were several large landmark studies that reported on the elevated risk of CKD during long-term PPI use. 13,14 In the study by Lazarus et al., 13 there were two large populations evaluated retrospectively. The findings suggested a hazard ratio of PPI use resulting in CKD of 1.17 and 1.5 in both groups compared with those not on PPIs. 13 With these data on hand, the absolute excess risk of PPI use causing CKD was calculated to be 0.1-0.3% per patient/year. 8 Xie et al. 14 found in a Veterans Affairs (VA) population that there was a hazard ratio of 1.28 in PPI users compared with the patients on H 2 RAs. Notably, both studies were retrospective in nature and come with inherent bias and possible confounding factors not accounted for in the analysis.
Another frequently reported adverse, long-term effect warranting more discussion is the possible link with dementia. Indeed, multiple studies suggested an elevated relative risk ranging from 4% to 80%. 8 The absolute excess risk was calculated to be 0.07-1.5%. 8 Multiple other retrospective analyses, along with a meta-analysis, have, however, suggested that there is no association between PPI use and progression to dementia. 10,15,16 The newest medication providing an alternative to PPI use, specifically in those with erosive esophagitis who would not have sufficient benefit from H 2 RAs or other milder treatments, is vonoprazan. This medication (a member of the new P-CAB class) works to block potassium channels at the H + K + -ATPase (i.e., the proton pump). 6 Thus far, this class of medications has not been linked to the adverse events reported in the long-term use of  18 Zhang et al., 19 and van Herwaarden et al. 20 ). * P < 0.05.
PPIs. That being said, this class of drugs is relatively new compared with the long-standing PPIs and it remains to be seen in postmarketing surveillance what potential downstream effects may be reported.

Conclusions
The take-home messages from the above data highlight the importance of discussing the potential risks and benefits of PPI use or nonuse. The ultimate goal should be to utilize these drugs in the appropriate patient for the appropriate indication. 8 If a PPI is absolutely indicated, which would be the case in PPI-responsive eosinophilic esophagitis (EoE) or erosive esophagitis, utilizing the lowest effective dose or switching to an H 2 blocker may be beneficial.
At the present time, PPIs appear to be one of the most effective medications practitioners have to manage acid reflux disease, the consequences thereof, or other conditions, such as EoE. A review of indication and assessment of potential risks is warranted when deciding if a patient should be prescribed any medication, including PPIs. Future therapies that may be available include the use of vonoprazan; however, this drug has not been submitted for approval in the United States but is currently undergoing evaluation for patients with erosive reflux disease.

Are there alternatives to PPIs in the management of GERD?
Since transient lower esophageal sphincter relaxations (TLESRs) are the main pathogenetic mechanism of GERD, 17 they represent a rational target for pharmacologic therapy. Baclofen, a gammaaminobutyric acid (GABA) receptor type B agonist, was the first reflux inhibitor identified. This drug decreased the number of reflux episodes per patient, the average duration of reflux episodes, and the incidence of TLESRs in both healthy volunteers and patients with GERD [18][19][20] (Fig. 1), a finding confirmed by a subsequent meta-analysis. 21 However, the incidence of adverse effects is not negligible, albeit not significantly higher compared with placebo. To improve the adverse effect profile, some reflux inhibitors (namely lesogaberan, arbaclofen placarbil, two GABA B agonists, and ADX10059, an mGlu5 antagonist) unable to cross the blood-brain barrier have been synthesized. Nevertheless, this class of drugs represents a promise unfulfilled, since the effect size was too small to be of clinical significance. 22 Patients with GERD can also meet the diagnostic criteria for another functional disorder, such as gastroparesis, since in some 30-40% of them, the gastric emptying (GE) of liquids and/or solids is delayed. 23 In this subgroup of patients, the esophageal exposure to acid is further enhanced thanks to the increased availability of gastric contents available for reflux. 24 Therefore, GI prokinetics would represent another pathophysiology-oriented therapy. 25 However, cisapride, the only prokinetic with well-documented efficacy in GERD, 26 has been withdrawn because of its intrinsic cardiotoxicity. 27 Metoclopramide and domperidone, besides having limited (if any) efficacy in GERD, display neurological 28 and cardiac 29 toxicity, and both the U.S. Food and Drug Administration and the European Medicines Agency have issued a black box in the summary of product characteristics (SmPC). Other compounds (like mosapride and itopride) are only available in Asian countries. The only available, safe prokinetic is prucalopride, a potent and selective 5-HT 4 agonist, currently approved for the treatment of chronic constipation, refractory to common laxatives. 30 This drug, which is devoid of cardiotoxicity, 31 is capable of accelerating GE and reducing the esophageal exposure to acid in healthy volunteers. 32 Despite the lack of specifically designed clinical trials, some case reports suggest its efficacy in controlling GERD symptoms refractory to PPI treatment. 33 Another 5HT 4 receptor agonist from the same chemical class, revexepride (at three different doses), failed to control regurgitation in patients with GERD symptoms partially responsive to PPIs. 34 In addition, the drug was no more effective than placebo at improving reflux parameters. 35 Acotiamide, an acetylcholinesterase inhibitor, which also enhances acetylcholine release, 36 was found capable of accelerating delayed GE and improving impaired gastric accommodation in patients with functional dyspepsia (FD), 37 as well as reducing the number of TLESRs in healthy subjects. 38 When added to acid suppressants, the drug improved symptoms in patients with PPIrefractory nonerosive reflux disease (NERD). 39 While prokinetics, added to H 2 RAs, improved symptoms and mucosal healing compared with the antisecretory drugs alone, this has been not the case when they were given with PPIs (for review, see Ref. 40). A meta-analysis of 12 randomized controlled trials, including 2403 patients, showed that despite partially improving patient quality of life, they have no significant effect on symptom improvement or endoscopic healing in patients with GERD. 41 A more recent systematic review, 42 assess-ing the potential benefits of mosapride (a selective 5-HT 4 agonist, with a major metabolite acting as the 5-HT 3 antagonist) added to PPIs, included seven trials with 587 patients. The results showed that the combination therapy is not more effective than PPI alone as the first-line therapy. Whether it is effective in PPI-resistant patients needs to be determined. On these grounds, prokinetics are currently not recommended by any guideline, even in patients with persistent symptoms on PPI therapy, 43 and should be used only on a case-by-case basis.

Alginate-containing formulations: merely barrier drugs or much more?
Alginate-containing formulations: mechanisms of action Alginate-containing formulations have long been used in GERD and stood the test of time. 44,45 However, advancements in the pathophysiology of reflux disease have prompted to give a new look to these "old drugs." 46 The advent of pH-impedance technology has represented a major advance, allowing us to understand that both acidic and weakly acid reflux are implicated in symptom generation 47 and that only those patients in whom acid is the symptom trigger respond to acid suppression. 48 On the contrary, alginate-containing formulations achieve symptom relief regardless of the stimulus (be it acid, pepsin, bile, or mixed). Their efficacy is likely due to the barrier effect, which translates into a reduction of the proximal migration of the refluxed gastric contents, 49 and binding and inactivation of pepsin. 50 Recent in vitro studies on esophageal biopsies from patients with endoscopy-negative reflux disease also showed a topical mucosal protective activity of alginate formulations against acidic and weakly acidic solutions. 51 Despite the buffering content of food, the pH of the material refluxed into the distal esophagus is very acidic owing to the presence of an acid pocket, which occurs in both healthy subjects and GERD patients. It represents an area of unbuffered gastric acid that accumulates in the proximal stomach after meals and serves as a reservoir for acid reflux. 52 The acid pocket, which is detected more frequently after large fatty or spicy meals, is longer in patients with GERD and in those with hiatal hernia (HH). 53 While in patients with a small HH, the acid pocket is located below the diaphragm, it is above the diaphragm in 40% of those patients with a large hernia. 53 3D echoplanar magnetic resonance imaging clearly shows, after the intake of an alginate-containing formulation (namely Gaviscon TM Advance), the development of a raft at the air-meal interface after a liquid meal and formation of a clot between the gastroesophageal junction and solid meal, exactly where the acid pocket would be located. 54 And indeed, in patients with GERD, the alginate-antacid raft colocalizes to the postprandial acid pocket 55 and displaces it below the diaphragm to reduce postprandial acid reflux 55,56 (Fig. 2). Because of the above pleiotropic actions, alginate formulations provide a symptom relief, which is not significantly different from that seen with antisecretory drugs 57 and, combined with PPIs, achieve a significantly higher therapeutic success. [58][59][60] Are the alginate-containing formulations all the same? It is worth emphasizing that the marketed alginate formulations contain a wide spectrum of alginatebased materials (soluble salts: sodium and potassium alginate; or insoluble salts: calcium and magnesium alginate) as well as different quantities of antacids (such as sodium bicarbonate and/or calcium carbonate). These formulations require three chemical reactions to take place simultaneously: transformation to alginic acid, sodium carbonate reacting with gastric acid to form carbon dioxide, and calcium salts releasing free calcium ions to bind with alginic acid, providing strength to raft formation. 46 Without gas production, alginates would mix with, and be emptied together with, food from the stomach. 61 As a consequence, the in vitro and in vivo behavior of each formulation may be different and clinical results obtained with one product cannot be extrapolated sic et simpliciter to the others. A recent comparative study evaluated the alginate content, raft characteristics, and acid-neutralizing capacity of seven different alginate-containing formulations in the UK market and found significant differences among floatation characteristics of the rafts. 62 A similar study performed on formulations available in Italy found markedly different weights and volumes (Fig. 3), as well as floatation characteristics, of the rafts generated by the different products. 63 In the past, even formulations of the same brand (namely Gaviscon) were found to possess significantly different raft strengths, 64 as they were produced by different local manufacturing plants, which is no longer the case today.
While alginate-containing formulations were prescription or OTC medicines, today many formulations are marketed either as medical devices, or even as dietary supplements. These new products have not been submitted to the same in vitro and in vivo scrutiny as the original formulations and cannot, therefore, guarantee the same pharmacologic activity and therapeutic efficacy.

Alginate-containing formulations for extradigestive GERD
Conversely from typical symptoms, the efficacy of PPIs on extraesophageal manifestations of GERD is uncertain. This uncertainty could result, at least in part, from available studies that are not homogenous, with differences in patient selection, endpoint considered, drug used, and regimen adopted. In addition, since extradigestive symptoms may need higher PPI dose and clinical improvement may take a longer time to occur, only properly designed trials would be able to unravel a clinical response. Unfortunately, however, this has not always been the case.
A careful analysis of the available literature shows that the efficacy of PPIs in extradigestive GERD is less consistent than that observed in patients with typical symptoms. In extradigestive GERD, the complexity of patient presentation is matched only by the challenge in making an appropriate diagnosis of reflux as the cause for the patients' complaints. Upper GI endoscopy and pH-impedance monitoring suffer from poor sensitivity, while laryngoscopy suffers from poor specificity in diagnosing reflux in this group of patients. 62,65 An empirical trial of PPIs could be the initial approach to diagnose and treat the potential underlying cause of these extradigestive symptoms; symptom resolution usually needs a higher PPI dose and longer treatment time than when PPIs are used in patients with typical GERD. 66 However, it is important to highlight that PPI therapy in extradigestive GERD and twice-daily dosing are nonapproved indications for these agents, although they are recommended by both GI 2,67 and other specialty guidelines. [68][69][70] For patients who improve with PPIs, GERD is presumed to be the most important etiological factor; but for those who do not respond, diagnostic testing with pH-impedance monitoring is reasonable to exclude continued or weakly acid reflux. Should this be the case, etiologies other than GERD should be considered. However, increasing evidence suggests that, in patients with proven GERD, PPIs alone may not suffice, and the use of add-on medications can achieve a higher success rate. Even with higher PPI doses and a longer duration of treatment, the response of patients with extradigestive symptoms can be substantially lower than that of patients with typical GERD symptoms.
In patients with laryngopharyngeal reflux, alginate-containing formulations achieve significant improvement in symptom scores and clinical findings, either alone 71 or in combination with PPIs. 72 Compared with acid suppression alone, the combination of esomeprazole and Gaviscon Advance attained a significantly better reduction of the reflux symptom index. 73 The efficacy of alginates in extraesophageal manifestations of GERD is likely due to its barrier effect, which translates into a reduction of the proximal migration of the refluxed gastric contents, 49 and binding and inactivation of pepsin. 50 The concentration and mucosal damaging activity of pepsin are potentially very high in the (acidic or nonacidic) refluxate that can reach the upper airways. 74 Is there a role for alginate-containing formulations in the long-term management of GERD? GERD is a chronic relapsing disease. Six months after discontinuing PPI therapy, 90% of patients with the erosive disease and 75% of those with NERD will relapse. 75 As a consequence, a longterm treatment (be it continuous, intermittent, or on-demand, according to the patient's clinical features) is often needed. 76 Dose titration may be necessary, allowing long-term maintenance therapy to be individualized. Indeed, the number and severity of relapses are highly variable among patients. Infrequent reflux symptoms are less likely to be chronic and may respond to different management strategies.
Given continuously or on-demand, PPIs are very effective in maintaining GERD patients in longterm (symptomatic and endoscopic) remission, 77,78 with the superiority of the full dose over the half dose. 79 In a large, multicenter French study, 80 an alginate-containing formulation (namely Gaviscon Advance) was given to control symptoms over a 6-month period to patients with reflux esophagitis, previously treated with H 2 RAs or omeprazole. On average, the symptom control was achieved in 76% of patients, with proportions depending on the initial severity of esophagitis. Among asymptomatic patients, 95% of them had taken only two alginate doses daily. 80

Conclusions
Alginate-containing formulations are an effective pharmacologic treatment of GERD. These formulations can be given either before (or instead of) PPIs in mild forms of the disease or in addition to PPIs either to manage breakthrough symptoms and/or to increase the success of antisecretory therapy. 81 Finally, alginate formulations can be given after PPI therapy to maintain symptomatic remission over time. In an era where the safety (especially longterm) of PPIs has been called into question (albeit not always correctly 82 ), 8,83 alginate-containing formulations are an effective and safe alternative.

Chemopreventive effects of PPIs in Barrett's esophagus
Biology and natural history A metaplastic change of the normal stratified squamous epithelium lining the esophagus into an intestinalized columnar epithelium, Barrett's esophagus (BE) is a complication of chronic GERD. The U.S. clinical guidelines recommend that, once BE develops, endoscopic surveillance be performed to monitor progression to dysplasia and esophageal adenocarcinoma (EAC), with the frequency of endoscopy depending on the presence and degree of dysplasia. 84,85 While various rates of progression of BE have been reported, it is now widely accepted that the annual rate is likely to be 0.22-0.33%. 86,87 Since esophageal cancer has a less than 20% 5-year overall survival rate, the goal of endoscopic surveillance in BE is to intervene with endoscopic eradication therapy before the development of invasive cancer. 88 An alternative strategy would be to prevent neoplastic progression of nondysplastic BE medically by interfering with the physiologic factors involved in progression, namely bile and acid reflux into the esophagus. As one example of interfering with progressions, a study in esophageal epithelial cell lines reported that the hydrophilic bile acid ursodeoxycholic acid (UDCA) protected against oxidative DNA damage and activation of inflammatory signaling pathways induced by the hydrophobic bile acid deoxycholic acid. 89 Following on this, in a surgically induced esophageal bile reflux rat model, the administration of UDCA significantly reduced the incidence of BE and EAC. 90 To determine whether these effects translated to patients, a pilot study was sponsored by the National Cancer Institute. Unfortunately, in 29 patients who took oral UDCA, markers of oxidative DNA damage, cell proliferation, and apoptosis (as assessed by immunohistochemical staining) did not change. 91 Gastric acid represents a genotoxic stress as it induces double-strand DNA breaks in BE cells and patient tissues. 92 Gastric acid can activate oncogenic signaling pathways, such as ERK and p38MAPK, and silence tumor suppressor genes, such as p16. 93,94 When combined with bile, gastric acid can also induce BE cells to undergo epithelial-to-mesenchymal transition, acquired by cells as they undergo transformation. 95 Thus, suppressing the effect of refluxed acid on BE cells should decrease their rate of neoplastic progression. While PPIs often effectively suppress gastric acid secretion at least enough to ameliorate heartburn symptoms, several studies have shown that they do not completely eliminate esophageal acid reflux in all patients with BE. 96,97 In a study of 48 patients with BE, half still had abnormally low esophageal pH despite reporting the absence of symptoms while receiving PPIs. 96 It has also been reported that reflux esophagitis heals more slowly in patients with BE, though the underlying cause remains unclear. 98

Chemoprevention by PPIs: evidence for and against
In 2014, a systematic review and meta-analysis of seven clinical studies (five cohort and two casecontrol) found a protective effect of PPIs on neoplastic progression of BE to high-grade dysplasia or EAC. [99][100][101][102][103][104][105][106] This meta-analysis included 2813 patients with BE (with or without low-grade dysplasia). PPIs were used by 84.4% of the patients (others were on H 2 RAs), with 317 progressing to low-grade dysplasia or EAC. Among patients, 75% were Caucasian, 82% were males, 22-58% were taking concomitant acetylsalicylic acid (ASA) or nonsteroidal anti-inflammatory drugs (NSAIDs), and 19-46% were on statins. In these patients, PPIs gave a 71% risk reduction in neoplastic progression. When separated by a duration of PPI exposure into short-(<2-3 years) or long-term (>2-3 years), only longterm PPI use was protective. Further, restricting the analysis to studies that were adjusted for the use of NSAIDs and statins demonstrated that the protective effect of PPIs was independent of these other medications. A more recent nested case-control study of 29,536 U.S. veterans matched 378 cases of EAC that developed in patients with BE to 1012 controls with BE without EAC. 107 Cases had a lower rate of PPI use and a shorter duration (<18 months) of therapy, while controls were on higher dose PPIs. Patients using PPIs had a 40% reduction in progression to esophageal cancer, while patients using H 2 RAs had a 30% reduction. A contradictory casecontrol study examined the rate of progression in BE patients in Denmark. 108 With a mean followup of 10.2 years, 140 cases developed either highgrade dysplasia or EAC. On the basis of prescription data, patients were also divided into low (<75%) or high (>75%) adherence to their PPI medication regimen. While low adherence patients had a 2.2 relative risk of progression, the high adherence patients had a higher relative risk of progression of 3.4. This was a somewhat surprising finding and the authors suggested that more severe reflux symptoms might drive higher compliance with PPI use. Adding this Danish study and another nested casecontrol study examining PPI use in the UK and the Netherlands to the seven studies in the 2014 metaanalysis, the authors of another systematic review with a meta-analysis came to the opposite conclusion that PPIs do not definitively protect against the development of either high-grade dysplasia or EAC. 109,110 Supporting the conclusions of this second meta-analysis, a nationwide study in Sweden of patients who took maintenance PPIs with a mean follow-up of 4.4 years showed that PPIs led to a higher risk of developing esophageal cancer, both adenocarcinoma and squamous cell carcinoma. 111 This finding was seen even in patients who took PPIs for clinical indications that were not associated with an increased risk for developing EAC (see Table 1, which includes Refs. 99-111).  106 Cohort 344 3.6 HGD/EAC de Jonge et al. 100 Case-control 335 N/A EAC Hillman et al. 103 Cohort 350 7.4 HGD/EAC Nguyen et al. 99 Case-control 812 N/A EAC Altawil et al. 102 Cohort 77 NR LGD/HGD/EAC Jung et al. 101 Cohort 355 5.9 HGD/EAC Tan et al. 107 Case-control 1098 N/A EAC Jankowski et al. 112 Randomized, factorial 2577 8.9 All-cause mortality/HGD/EAC Singh et al. 105 Meta-analysis 2813 HGD/EAC Opposing Hvid-Jensen et al. 108 Case-control 1437 10.2 HGD/EAC Masclee et al. 109 Case-control 1466 NR HGD/EAC Brusselaers et al. 111 Cohort 6040 4.4 EAC Hu et al. 110 Meta-analysis 5712 HGD/EAC EAC, esophageal adenocarcinoma; HGD, high-grade dysplasia; LGD, low-grade dysplasia; N/A, not applicable; NR, not reported.
On the basis of contradictory results from these prior observational studies, the results of the first phase III randomized, prospective trial examining the chemopreventive effects of PPIs on progression in patients with BE were eagerly awaited. 112 The trial also examined the chemopreventive effects of ASA, on the basis of its activity inhibiting the inflammation-related enzyme cyclooxygenase-2. The Aspirin and Esomeprazole Chemoprevention in Barrett's Metaplasia Trial (AspECT) randomized 2557 patients with BE to receive a high-dose PPI (esomeprazole 40 mg twice-daily), low-dose PPI (esomeprazole 20 mg once-daily), high-dose PPI with ASA (300-325 mg ASA), or low-dose PPI with ASA. Inclusion criteria included an age of at least 18 years and endoscopic evidence of at least 1 cm of histologically proven BE, while exclusion criteria included existing high-grade dysplasia, EAC, or NSAID use at baseline. Patients were reevaluated annually for up to 10 years with a median follow-up period of 8.9 years. During odd years, patients were evaluated in the clinic and questioned about hospitalizations and serious adverse events. During even years, they underwent endoscopic screening of the esophagus with four quadrant biopsies every 2 cm of Barrett's metaplasia. Histopathologic development of either high-grade dysplasia or EAC was confirmed by two pathologists. The composite primary endpoint was all-cause mortality, high-grade dysplasia, or EAC, whichever occurred first. Intention to treat analysis and accelerated failure time modeling were used to assess patient outcomes. Because some patients could not tolerate ASA or were taking lower doses of ASA for cardiovascular indications, these patients were randomized to one of the two PPI only groups, increasing the numbers in these two groups. The high-dose PPI significantly increased the time to reach the outcome events over the low-dose PPI (Fig. 4). The effect of ASA was not significant; however, if patients who initiated NSAID use during the trial were removed from analysis at the time point of the first use of NSAIDs, ASA gained a significant protective effect. This was borne out when the largest difference in time to the outcome was found between high-dose PPI and ASA and low-dose PPI and no ASA. When secondary analyses for individual components of the composite endpoint were performed, high-dose PPI decreased all-cause mortality compared with low-dose PPI but did not significantly alter time to develop high-grade dysplasia or EAC. 112 On the basis of the results of this prospective randomized trial, it is appropriate to continue using PPIs as chemopreventive agents in patients with BE until more precise biomarkers define which patients are at the highest risk of neoplastic progression and, therefore, should be placed on high-dose PPI therapy. 112

Nonantisecretory effects of PPIs
Despite their concentration within the secretory canaliculi of the parietal cell and their pharmacologic selectivity, 113 PPIs also have some side effects that are often (but not always) independent from their antisecretory activity (Table 2). Sir William Osler once commented that no drug has a single effect 114 and PPIs are no exception.

Effects of PPIs on gastric motility
In the upper and lower GI tracts, gastric secretion and motility are two interrelated functions that cannot be regarded or studied independently from each other. 115 From a physiological standpoint, it is well known that the main stimulants of acid secretion (namely acetylcholine, histamine, and gastrin) also modify GE by a mechanism that may be independent of their secretory effect. 116 As a consequence, antisecretory drugs often affect gastric motility, while motor-stimulating compounds seldom modify the secretory pattern. 117 With one exception, all the studies on GE were performed on healthy volunteers by using different techniques to measure GE (e.g., gastric intubation, scintigraphy, ultrasonography, and paracetamol absorption kinetics). The different PPIs were given by the oral route in single or repeated doses (for review, see Refs. 118 and 119). The majority of studies, which have been performed with omeprazole, clearly show that this PPI is able to delay the GE of both liquids and solids. The delay in GE was evident at both the onset of GE (as evidenced by prolongation of the lag phase) and during the linear emptying (as shown by the slope of the emptying curve). Compared with the omeprazole results, data obtained with lansoprazole and rabeprazole were less consistent but nevertheless showed that both antisecretory compounds can, under given experimental conditions, delay GE, especially of solid food. 118,119 The mechanism(s) by which omeprazole and other PPIs delay the emptying rate has not been fully elucidated, as yet. The pivotal role that acid and pepsin exert during the process of triturition 120,121 and the significant correlation between inhibition of GE and changes in postprandial pH under omeprazole treatment 122 strongly suggest the motor effect of PPIs is mainly due to inhibition of gastric acid secretion. This will result in inadequate hydrolysis of food leading to gastric persistence of particles that are too large to cross the pyloric sphincter. This delay is evidenced in scintigraphic studies by prolongation of the lag phase. 123 An interesting experiment, performed in chickens, 124 showed that adding hydrochloric acid to food significantly reduced omeprazole-induced delay in GE, thus confirming that this peculiar pharmacological activity of substituted benzimidazoles is acid mediated.
The antisecretory effect of PPIs also translates into a significant decrease of intragastric fluid secretion leading to increased viscosity, mainly of the liquid phase of gastric contents. This will reduce Table 2. Nonantisecretory effects of PPIs r Inhibitory activity on GI motility r Mucosal protective activity r Antioxidant activity r Anti-inflammatory activity r Antibacterial activity r Antineoplastic activity discrimination between solid and liquid emptying 125 and result in a prolongation of the emptying rate. 126 Finally, PPI-induced gastrin release 127,128 may also have a role in the delay of GE. The peptide was indeed shown capable of either stimulating antral contractility or delaying the emptying rate, 129 actions both mimicked by omeprazole. 121 All these events may contribute to triggering the observed gastric motor effects of PPIs.
As a consequence of drug-induced gastric motor derangement, dyspeptic symptoms may actually be worsened by PPI therapy or, alternatively, new symptoms (especially postprandial fullness) may arise during treatment. This could be more relevant in GERD patients with delayed GE 130 or with associated FD. 131 If this is the case, patients should be switched to H 2 RAs, ranitidine, or nizatidine, which, in addition to their antisecretory effect, display a cholinergic-like activity 118 and have been shown to accelerate GE. 132 In this regard, a Cochrane metaanalysis 133 concluded that H 2 RAs are better than placebo in achieving symptom relief in patients with FD.
Mucosal protective activity of PPIs PPI administration represents the standard of care, recommended by International guidelines, 78,134 to prevent NSAID-associated gastroduodenal damage. They are more effective than H 2 RAs, which can prevent duodenal but not gastric damage.
In comparative studies, the antisecretory and gastroprotective effects of omeprazole were generally equal to, or exceeded, those of H 2 RAs in different (not always acid related) experimental models of gastric injury, suggesting a mucosal protective activity, independent from the antisecretory effect. 135 Several investigations have shown that PPIs are able to increase mucus secretion, via stimulation of endogenous prostaglandin release and activation of mucosal sulphydryl radicals as well as to inhibit oxygen-derived free radical production by human neutrophils. 135 Taking into account that oxygenfree radicals play a pivotal role in the development of NSAID-induced mucosal lesions, 136 this pharmacological action could also contribute to the protective effect of PPIs on the gastroduodenal mucosa.
Besides preventing NSAID-induced gastroduodenal ulcers, PPIs are also effective in healing them. 78,134 It is well known that NSAIDs inhibit cell proliferation in the gastric mucosa at the ulcer margin both in rats and humans. This effect may account for the epidemiologic findings of delayed ulcer healing in patients taking long-term NSAID therapy. 136 The mechanisms underlying the therapeutic effect of PPIs involve a decreased acid-peptic attack at the ulcer base as well as a reduced degradation of acid unstable fibroblast growth factors (FGFs). 137 Ulcer healing is a complex process, which is initiated by the secretion of FGFs in the ulcer base and margin. These factors promote the proliferation of epithelial cells and their margination over the ulcer crater. During healing, granulation tissue undergoes continuous remodeling and is usually transformed into a thinner mature scar within 2 weeks. The stimulation of angiogenesis, which has the potential of supplying oxygen and nutrients to the granulation tissue, also represents a major mechanism of acceleration of ulcer healing. 137 Experimental studies have shown that indomethacin decreases angiogenesis and increases the thickness of the granulation tissue, the effects both counterbalanced by omeprazole administration. 138,139 The esophageal mucosal protective activity of PPIs was studied by investigating their effect on the mucosal barrier in patients with PPI-responsive esophageal eosinophilia (PPI-REE) or EoE. 140 The integrity of the esophageal mucosa is impaired in both these patients, allowing transepithelial transport of small molecules and allergens. Indeed, in vitro studies from esophageal biopsies found that transepithelial electrical resistance is reduced and intercellular spaces are dilated. PPI therapy partially restores mucosal integrity in patients with PPI-REE, but not in those with EoE. 140 This normalization of dilated intercellular spaces is similar to that observed in patients with GERD. 141 The clinical relevance of PPI mucosal protective activity is difficult to ascertain. Indeed, according to the experimental studies, doses higher than the antisecretory ones are needed to exploit it. However, mucosal protection could be of value, in addition to acid suppression and to the anti-inflammatory activity (see below), in healing reflux esophagitis 77 and EoE 142 as well as in the prevention of NSAID gastropathy. 134,136 Antioxidant and anti-inflammatory activities of PPIs A recent in vitro study 143 compared the antioxidant activity of different PPIs, by measuring their ability to scavenge 2,2,1-diphenyl-1-picrylhydrazyl (DPPH) free radicals in comparison with a standard antioxidant substance (namely vitamin C). All the studied drugs displayed an antioxidant activity but to a different extent. Omeprazole and esomeprazole showed the highest ability to scavenge free radicals with a potency similar to that of vitamin C. 143 This study and a previous one 144 revealed that the antioxidant effect of PPIs can be mediated by the parent drugs, and conversion of these drugs to their protonated form retains their antioxidant activity. PPIs can directly bind to neutrophils and attenuate oxygen-derived free radical production and, in a model of ethanol-induced gastric damage, are able to inhibit the expression of adhesion molecules on neutrophils and vascular endothelial cells, thus providing a mechanism for the suppression of inflammatory cell activation. 145 The antioxidant activity of PPIs has also been observed in several in vivo studies at doses much lower than those needed to inhibit gastric acid secretion (for review, see Ref. 146).
Besides the above described antioxidant effect, there are several other potential mechanisms underlying the anti-inflammatory effects of PPIs (Table 3). These include effects on inflammatory cells, endothelial and epithelial cells, as well as on gut microorganisms, 146 including Helicobacter pylori, Plasmodium falciparum, Leishmania donovani, and some yeast. 147 It is not clear whether oral PPI dosing can achieve the high drug concentrations in plasma and tissue that would be needed to reproduce some of the anti-inflammatory actions, observed in vitro and in vivo experimental settings. Nevertheless, both antioxidant and anti-inflammatory properties of PPIs may contribute substantially to their efficacy in EoE. As a matter of fact, in vitro and in vivo studies suggest that the anti-inflammatory effects of PPI therapy rather than acid suppression alone may be  responsible for the observed clinical and histologic improvement through inhibition of the T H 2 allergic pathway. Indeed, like topical corticosteroids, PPIs downregulated cytokine expression. 148 PPI therapy significantly downregulated the esophageal eotaxin-3/T H 2-cytokine gene expression in PPI-REE, similar to that seen in a steroid-responsive EoE. 148 Antineoplastic effects of PPIs Besides some potential and alleged indications of PPIs in cancer patients (for review, see Ref. 78), increasing evidence suggests, for this class of drugs, an antitumor effect (through the selective induction of apoptosis as well as an anti-inflammatory effect). 149 They also exert protection of cancer cells from developing chemo-or radiotherapeutic resistance. 150 Acidification of the extracellular compartment represents a conceivable mechanism of drug resistance in malignant cells. In addition, it drives proliferation and promotes invasion and metastasis. 150 Experimental evidence has shown that PPIs counteract tumor acidification (via inhibition of the vacuolar H + -ATPase) and restore sensitivity to anticancer drugs. Moreover, early clinical data have supported their role as add-on medications to anticancer treatments in patients with osteosarcoma or breast cancer. 151 A large epidemiological study (the SPORE program 152 ) found that patients with head and neck tumors, taking acid suppressants, had significantly longer overall survival compared with those who did not. In a case series, patients with lower GI cancers, treated with an unconventional association of high-dose PPIs and metronomic chemotherapy, had a good clinical outcome, despite the failure of previous standard chemotherapy for metastatic disease. 153 On these grounds, a randomized phase II trial to evaluate the combination of PPIs and metronomic capecitabine, as salvage treatment for patients with advanced GI tumors, is ongoing. 154 Specifically designed clinical trials to better characterize the role of PPIs as new therapeutic agents in cancer treatment are underway. Therefore, PPI use as an adjunct to cancer chemotherapy should not be performed outside a clinical trial.

Competing interests
C.S. has served as a speaker, consultant, and/or advisory board member for Alfasigma, Pfizer, Takeda, Reckitt-Benkiser, and Shionogi, and has in the past received funding from Giuliani Pharmaceuticals and Pfizer. R.H.H. has served as a speaker, consultant, and an advisory board member for AstraZeneca, Danone, GSK, Merck, Pfizer, and Takeda. J.A.S. and D.H.W. declare no competing interests.