The widespread use of proton pump inhibitor drugs (PPIs) world wide for almost two decades is gradually increasing concern among physicians and the public that their benefits may be attended by a variety of risks that until recently have received little attention. Based largely on observational studies, a number of adverse drug reactions (ADRs) have emerged as likely or possibly associated with PPI therapy. Such ADRs are either uncommon in incidence, or associated only with prolonged use or high doses of the drugs. Not all these ADRs are class effects, although often erroneously treated as such, and some may occur only in subpopulations, susceptible because of age, genetics, comorbidity or environmental factors such as Helicobacter pylori infection or cotherapy with other agents.
The incidences of minor adverse events, identified in premarketing clinical trials, are typically of the order of 1–5% and include headache, diarrhea, constipation, nausea and rash, with little difference between products and with only minor differences from those seen with H2-receptor antagonists (H2RAs) or placebo. These effects of therapy will not be discussed here except to say that the short-term use of the drugs in recommended doses is generally well tolerated, even in children, the fetus, the pregnant and the elderly, with only rare exceptions. Nonetheless, from studies not designed to determine drug safety systematically, a number of associated ADRs have emerged and are the focus of this brief commentary. Due to limitations of space, this review will not attempt to discuss all relevant studies or provide detailed bibliographic references: for more information on specific points, the interested reader is referred to other recent reviews [1,2•–4•].
The available literature mainly identifies associations of PPI therapy, often weak, and frequently accompanied by unsupported speculations as to the mechanisms that underlie them. However, such observational data, despite the listed imperfections, are often later borne out by the results of randomized controlled trials (RCTs), and prove to have furnished valuable clues to important effects, unrecognized prior to extensive postmarketing use. That these suspected ADRs are unproven in no way justifies the belief that PPI therapy is completely well tolerated. Most prominent among associations of ADRs with long-term PPI therapy are concerns about the risks of bone fractures, increased susceptibility to infections and effects of altered gastric function.
Effects on bone
To date, four studies have shown an association between chronic PPI use and fractures of the hip or other sites, including spine, wrists and forearms [5–8], the risks increasing (in two reports) with dose of drug and duration of exposure. Most of the reported hazard ratios or odds ratios (ORs), for various sites in various groups, were between 1.0 and 2.0, except for an OR of 2.65 [confidence interval (CI) 1.80–3.90] attending the use of more than 1.75 daily doses per day for more than 1 year. These risk increases were applied to low baseline risk rates in untreated controls. Some have used the term ‘osteoporotic fractures’ to describe these events, but as at least three studies have later indicated, PPI therapy is not associated with any appreciable changes in bone density, over periods adequate to reveal the effects of other relevant factors, such as steroids [8,9•]: the term osteoporotic, with implications as to pathogenesis, should now be dropped. Similarly, attempts to link hypochlorhydria with significantly impaired calcium absorption, have failed to show any consistent relationship [4•,9•], although other hypochlorhydric states, such as pernicious anemia or following gastrectomy, have also been associated with increased rates of hip fracture .
Two studies were unable to find an increased risk of hip fracture, unless at least one additional risk factor such as steroid use or preexisting osteoporosis was present [11,12••], suggesting that PPI use may be dangerous mainly in vulnerable subgroups. The authors involved in these studies emphasize their awareness that unrecognized confounding may play a major role in associating fractures with PPI therapy. Although the US FDA have recently issued some cautionary advice about the use of long-term or high-dose PPI therapy, the authors of the report also felt that it was not clear that PPI use accounted for the increases in fractures seen in PPI users in epidemiological studies.
Despite all these reservations and failures to demonstrate either drug-related calcium malabsorption or decreases in bone density, it seems likely that in some vulnerable individuals, PPI use contributes to an increased risk of fractures at various sites, in a manner related to dose of drug and duration of therapy. Although the mechanism(s) underlying their genesis remains unknown, several possibilities are worthy of examination. Osteoclasts in bone contain H+/K+ ATP-ases that acidify vacuoles in bone matrix during vacuolar resorption: two studies have shown that omeprazole inhibits bone resorption in vitro and in vivo in humans [2•,4•], decreasing bone turnover. This should lead to increases in bone density as measured by bone scans but, as pointed out by Targownik et al. [9•], might increase fracture risk by blocking the repair of microfractures, thus weakening bone strength. They observed that in the rare disease osteopetrosis, in which vacuolar osteoclastic proton pumps are congenitally absent, bone remodeling is impaired, bone density is increased and repeated fractures occur [9•].
Another factor might be deficiency of vitamin B12, whose absorption is impaired in some PPI users and which is integral to bone development, especially to osteoblastic function: deficiency of B12 has been associated with increased risk of fractures . In all, but especially in H. pylori-infected patients, PPI use dose dependently leads to hypergastrinemia and in animal models, leads to hyperparathyroidism and increased bone turnover. On the contrary, PPIs appear to reduce the absorption of magnesium, causing hypomagnesemia, which, in turn, inhibits parathyroid function and may cause tetany. There have now been four reports of hypomagnesemia caused by PPIs, due to impaired absorption of magnesium. This could result from actions of PPIs that interfere with the TRPM6 and TRPM7 ion channels that conduct divalent cations into cells [3•]. These channels are regulated strongly by protons in several sites, and PPI-induced alterations in gut pH could thus influence absorption of magnesium or calcium, in turn affecting parathyroid function, bone metabolism and susceptibility to fracture.
There is an urgent need for careful prospective studies of the effects of PPIs on bone metabolism and for, epidemiologic studies carefully designed to minimize confounding by various clinical factors. Meanwhile, the overall increase in fracture risk appears small. Although the increased risk does not require discontinuation of needed PPI therapy, high-dose or long-term use should be confined to situations in which any increase in risk is clearly justified by the need for PPI in the particular case. A similar risk, though of lesser magnitude, probably also applies to the use of H2-receptor antagonists (H2RAs) [5,12••]. Much of the fracture risk may ultimately prove to arise from widespread use of PPIs in very sick people whose comorbid illnesses collectively increase the risk of fracture.
Susceptibility to infection
The principal function of gastric acid in humans is, as a host defense, to sterilize all contents entering the digestive tract. It is effective against sensitive organisms, preventing infection unless the inoculum is large, acid-resistant, highly virulent or host defenses are impaired. Reduction in acidity is associated with increased risks of both enteric and systemic infection. Following the famous admonition of Sir Arthur Hurst to the British Government in 1934, that people with hypochlorhydria should not be sent to the tropics, there have been numerous previous papers on the risks of infection associated with hypochlorhydria caused by disease, surgery or therapeutic agents .
In a recent review, Dial [14••] has analyzed effects of PPIs on gastric host defense against microbes. Acid secretion is believed to be of key importance in killing organisms, but elevations in gastric pH, delayed gastric emptying, increased bacterial translocation, decreased gastric mucus viscosity and changes in flora may all contribute. The only organism uniquely adapted to surviving the acid milieu is H. pylori, and the effects of PPIs in inhibiting gastric acidity/elevating pH are significantly greater in H. pylori-infected individuals than in controls. In the absence of gastritis, gastric acidity does not decrease with age: the increasing prevalence of hypochlorhydria or achlorhydria in elderly individuals is mainly due to chronic gastritis and increasing atrophy and relates to a birth-cohort effect associated with the risk of H. pylori infection in the individual's youth, and not to age per se . H. pylori gastritis is likely to be an important confounding factor in our understanding of an apparent increase in risk and severity of C. difficle infections in the elderly, especially when treated with acid suppressive drugs. In trials, all marketed PPIs cause some diarrhea, but how much of this is of infectious cause has been little studied. Some diarrhea may also be caused by small bowel bacterial overgrowth (SIBO) as discussed later. However, because of the rising incidence, costs and severity of the illness, much scrutiny has been devoted to the risk of Clostridium difficile-associated diarrhea (CDAD).
Clostridium difficile and other enteric infections
As borne out by systematic analysis  and as reviewed by Dial [14••], there is a consistent association between PPI therapy and an increased risk of CDAD, with a pooled OR of 1.94 (95% CI 1.37–2.75) for any antisecretory therapy, of 1.96 (95% CI 1.28–3.0) for PPI therapy and 1.4 (95% CI 0.85–2.29) for H2RA therapy. The OR for Salmonella, Shigella or other enteric infection was 2.55 (95% CI 1.53–2.46). Similar or greater increases in C. difficile colitis have been observed in patients with chronic renal failure, in high-risk and average-risk hospital settings, and in community studies. In a large pharmaco-epidemiologic cohort study, involving 101 796 hospital discharges from a tertiary care center over 5 years , the occurrence of nosocomial CDAD increased from 0.3% (95% CI 0.21–0.31) in patients receiving no acid suppression, to 0.6% in those receiving H2RAs (95% CI 0.49–0.79), to 0.9% (95% CI 0.8–0.98) in those receiving once daily PPI and to 1.4% (95% CI 1.15–1.72) in those receiving more frequent PPI therapy. After adjustments for age, comorbid conditions, antibiotic exposure and other factors, ORs rose from 1.53 (H2RAs), to 1.74 (daily PPI), to 2.36 (>daily PPI). These figures support a dose–response effect and argue for causality. Finally, the risk for recurrent CDAD during 90 days after discharge from hospital was increased 42% by cotherapy with PPIs during attempts to abolish CDAD . These results are impressive but still need to be confirmed in prospective, randomized studies.
In recent times, we have come to realize that C. difficile spores are frequently spread between patients, particularly in hospitals, and cause illness even in the absence of prior antibiotic use: normal enteric flora protect against their activation and greatly diminish the risk of infection. Antibiotic therapy abolishes this protection. In attempts to understand the effects of PPIs on CDAD, attention has focused on the pH-dependent conversion at higher pH of the acid-resistant spores to the more virulent vegetative form of the organism, by delayed gastric emptying prolonging the duration of this process, and possibly by effects of bile salts, potassium and phosphate in gastric juice [14••]. An accentuation of these effects by H. pylori infection has not been studied, but may account for the increased risk in the elderly. The marked increase in CDAD infection in the past two decades has coincided in time with increasing use of PPIs, but other factors such as the emergence of more virulent strains may also play a part.
In conclusion, it appears highly likely that PPI therapy will prove to increase the risk of infection with a number of gastrointestinal pathogens, among which C. difficile is the most important, and most in need of prospective studies. There is increasing need to consider possible risks and benefits before prescribing or continuing PPI therapy, particularly in hospitalized patients on antibiotics or during institutional outbreaks, in the frail, the elderly or the immunosuppressed, and in those embarking on travel to areas of risk.
Although a number of early studies suggested a weak association between community-acquired pneumonia (CAP) and PPI use, a recent survey of 70 RCTs, of which seven met strict criteria for inclusion into meta-analysis, found no association, although about half the studies showed a trend toward significant increases. In two separate studies, there appeared to be a high risk of CAP in the first 5–14 days of therapy, but nothing for longer times or with chronic exposure. There are no convincing data to indicate significant links with nosocomial or ventilator-associated pneumonias . Despite abnormal gastric colonization in PPI users, and theoretical risks of increased micro-aspiration or translocation, no reports have appeared. Reports of increased leakage of the gastric mucosa (probably via tight junctions) caused by several PPIs so far apply only to molecules in the 500–4000 Da weight range and not to larger molecules or particles, but this may change : this phenomenon is unlikely to affect bacteria.
Small intestinal bacterial overgrowth
Despite long-standing suspicions that acid-suppressive therapy increases the risk of SIBO, it has been uncertain whether this risk is uniform in all users or is perhaps restricted to certain subgroups, such as the elderly, the H. pylori-infected or those with irritable bowel syndrome (IBS). There is growing acceptance of the presence of SIBO in 25–40% of patients with IBS-type symptoms, particularly gaseous bloating, excessive flatus and diarrhea, but also to a lesser extent in those in whom constipation or pain are dominant symptoms. Whether PPI use is an independent risk factor for SIBO in individuals free of IBS is not well studied. Furthermore, the type of overgrowth due to PPIs may be caused more by oropharyngeal flora rather than by the usual flora of the more distal gut and give rise to a different clinical picture. Despite such reservations, in a recent study of 450 patients with a positive glucose hydrogen breath test (GHBT), among 200 GERD patients on PPIs, GHBT was positive in 50% compared with 24.5% among 200 PPI nonusers with IBS and 6% among 50 normal controls, all differences between groups being highly significant . Thus, even if there were IBS patients among the PPI users (unstated), the combination of IBS and PPI therapy carries at least twice the risk seen with IBS alone: an accompanying editorial identifies some additional reservations.
Although some other recent studies have not confirmed the observation, their designs [small numbers, use of lactulose breath test (LBT), etc.] do not allow confident rejection of the reported association. Furthermore, the association is strongly supported by a number of other observations, including chronic use of PPIs as an independent risk factor for SIBO in the elderly (OR 7.9), PPI therapy increasing the occurrence of relapses of SIBO after eradication by antibiotic therapy (OR 3.52) and PPI use an independent predictor of positive GHBT in 2000 non-IBS patients with a variety of diseases such as diabetes, cirrhosis, collagen vascular diseases or residua of previous gastric surgery (OR 1.27, P = 0.028). Older studies, culturing small bowel bacterial contents, also found increased risks of SIBO with H2RA therapy, greater for ranitidine than for cimetidine, and that the incidence of SIBO rose in scleroderma patients changed from H2RAs to PPIs.
In conclusion, although there is an undoubted need for randomized prospective studies in this area, it seems likely that chronic use of PPIs increases the risk of SIBO in a sizable number of individuals, the risk increasing with time on drug. This number may be influenced by a number of other factors including the presence of IBS, chronic gastritis, H. pylori infection, other comorbid conditions and exposure to antibiotics or to drugs or diseases that affect gastric emptying.
Spontaneous bacterial peritonitis
Although the pathogenesis of spontaneous bacterial peritonitis (SBP) is incompletely understood, translocation of bacteria across the intestinal wall is believed to be an important causative factor. It follows that alteration in intestinal flora, such as occurs with PPI use, may play an important role in SBP, although additional factors are clearly involved. It has been hypothesized that PPIs, widely used in cirrhotic patients, may favor intestinal bacterial overgrowth and translocation, increasing the risk of SBP. One recent retrospective case–control study of 70 cirrhotic, community-based patients with SBP, on multivariate analysis found that PPI use was associated with SBP with an OR of 4.31 (95% CI 1.34–11.7). Affected patients had advanced cirrhosis, and higher ascitic fluid protein was protective. Of interest was that 47% of patients had no indication for PPI treatment .
Inhibition of gastric secretion of acid, pepsin, intrinsic factor, vitamin C and other substances has given rise to concerns about a number of possibly resulting clinical deficiency states.
Both hydrochloric and ascorbic acids secreted by the stomach are involved in converting ferric to ferrous iron and maintaining it in the reduced state. Ascorbic acid, whether of dietary or gastric origin, also chelates nonheme iron and keeps it in solution until absorbed. PPIs inhibit secretion of both substances. In addition, vitamin C is unstable at high pH and is readily converted to dehydro-ascorbic acid in inflamed H. pylori-infected mucosa. Thus, conditions for iron absorption are significantly suboptimal during PPI therapy, particularly in the presence of H. pylori gastritis. Despite use of PPIs to reduce iron absorption in hemochromatosis and observations that omeprazole inhibited the correction of iron deficiency anemia by oral iron until the PPI was stopped, there has been no prospective study of the association of PPI therapy with iron deficiency anemia and no proper trial of the efficacy of oral iron administration during PPI therapy . PPI use has also been linked with the possible development of painful restless legs syndrome, long associated with iron deficiency and low serum ferritin concentrations. In my opinion, iron deficiency is the most important of the uninvestigated potential adverse effects of PPI therapy, especially in the elderly and the poorly nourished, and is in urgent need of prospective studies.
PPI therapy leads to diminished acid secretion, diminished antral D-cell release of somatostatin, consequent increased G-cell release of gastrin and hypergastrinemia. This causes oxyntic cell hyperplasia, increased parietal cell mass, glandular dilatations and stimulation of enterochromaffin-like (ECL) cells to release chromogranin and histamine, raising their concentrations in serum. Cellular hyperplasias can cause obstructions at the mouths of fundic glands, causing them to balloon, become prominent and polypoid in appearance. Hyperplastic or cystic-type fundic polyps arise in 7–10% of the patients taking PPIs for 12 months or more: these are benign and regress with cessation of therapy. However, in patients with familial adenomatous polyposis, fundic polyps may progress during PPI therapy, becoming dysplastic and harboring adenocarcinoma: these should be excised during surveillance endoscopy [4•]. Other specific consequences of hypergastrinemia are now briefly reviewed.
Rebound acid hypersecretion
This is an effect of acid suppression recognized for many years by basic investigators but largely ignored by clinicians [24••,25••]. Sustained hypergastrinemia due to daily PPI therapy leads to increased gastric acid-secretory capacity that is not apparent during PPI therapy, but that appears promptly when the drug is stopped. In normal volunteers, treated for 8 weeks with esomeprazole 40 mg per day, stopping the drug was followed by the de-novo development of dyspeptic symptoms in more than 40% of previously symptomless patients. Other smaller studies, published as abstracts, had previously reported similar increases in acid-related symptoms after stopping PPI therapy, both in normal volunteers and in symptomatic patients, and are discussed elsewhere. Depending on the dose and duration of exposure, it can take 2–3 months for rebound acid hypersecretion (RAHS) to return to pre-PPI basal levels [24••]. Like nocturnal acid breakthrough, RAHS is not prominent in H. pylori-infected patients.
This phenomenon implies that a post-PPI-therapy period of difficulty should be anticipated. Symptoms can be managed with antacids or H2RA drugs (which cause negligible RAHS and can be discontinued later with little difficulty), or perhaps the PPI dose could be tapered down over a period rather than suddenly discontinued: to date, too few studies of tapering drug cessation have been reported for any conclusion to be drawn. Furthermore, as gastrin acts by releasing histamine, the ultimate mediator of acid secretion, tolerance to histamine and to the effects of H2RAs occurs during PPI therapy, decreasing H2RA efficacy; thus, PPIs should not be used in advance of H2RAs, an argument against ‘step-down’ therapy and other popular practices [24••].
Impaired gastric emptying of solids
This action can be of importance in managing functional dyspepsia and gastroesophageal reflux, particularly in those whose emptying is already delayed due to diabetes, vagotomy or the use of opiate drugs. The volume of gastric contents is important not only for patient comfort but also for triggering transient lower esophageal sphincter relaxations (TLESRs) and for exacerbating reflux in those with low sphincter pressures, particularly at night and in recumbency. A detailed scholarly review of the physiology of gastric emptying and its alteration by PPIs has recently been published by investigators from Japan [26•].
Hypergastrinemia and neoplasia
Gastrin has trophic effects on many tissues and stimulates a number of tumor cell lines in culture, including colon cancer cells. Although there have been suggestions that hypergastrinemia is associated with an increased risk of colon cancer, several recent large studies have found no such increase in patients using PPIs [2•]. Neither does PPI therapy increase the risk of adenocarcinoma of the stomach or cause progression of gastritis or gastric atrophy, except possibly in patients with active H. pylori infection .
More troubling, because of the extensive use of PPIs in patients with GERD, is a possible effect of gastrin on the progression of Barrett's esophagus to cancer. This is the subject of much recent research and controversy: the bulk of the data seem to suggest that in most patients progression of dysplasia is halted or reversed by PPI therapy . However, in a recent study, although there was no significant difference in serum gastrin with increased degrees of neoplasia, in multivariable analysis, the highest quartile of serum gastrins was associated with a significantly increased OR for advanced neoplasia (high-grade dysplasia or adenocarcinoma) with an OR 5.46 (95% CI 1.20–24.8). This finding suggested that some subset of Barrett's esophagus patients may be cancer prone, perhaps responding only to serum gastrin above a certain ‘threshold’ level . The issue remains unresolved, but of considerable interest, in view of the marked rise in the incidence of adenocarcinoma at the cardioesophageal junction over the past two decades, as acid-suppressive therapy for GERD has greatly increased.
PPIs inhibit acid secretion, leading antral G cells to release gastrin, causing hypergastrinemia. Gastrin, in turn, binds to gastric mucosal ECL cells, causing them to release chromogranin, histamine and other substances. The acid-secretory effects of gastrin are inhibited by PPI, but the potential proliferative effects on mucosal cells or cancers are not. In most patients, PPI-induced elevations in serum gastrin are moderate (50–400 pg/ml) and normalize when the drug is stopped. However, in some patients (mostly those with H. pylori infection and atrophic gastritis), plasma levels can rise to between 400 and 4000 pg/ml. The trophic, concentration-dependent effects of gastrin are exerted at much lower concentrations, above 40 pg/ml.
In rodents given PPIs, hypergastrinemia leads to the development of gastric carcinoids; hypergastrinemia also does this in both the genetically prone MEN-1 subset of humans with Zollinger–Ellison syndrome (hyperchlorhydric) and in another genetic disorder pernicious anemia (hypochlorhydric). In GERD patients, long-term PPI use is associated with the development of focal areas of hyperplasia of ECL cells in 10–30% of patients: the genetic contribution to this is unknown. Whether or not PPI therapy leads to the development of gastric or other carcinoids in humans has never been formally studied, despite large recent increases in the incidence of these comparatively rare tumors in many sites. These increases have occurred over the past three decades . The annual population incidence in gastric carcinoids has risen from 0.03% (1969–1971) to 0.12% (1992–1999) in men and from 0.02% (1969–71) to 0.18% (1992–1999) in women, increases of 400 and 900%, respectively, and the percentage of all carcinoids occurring in the stomach has also risen from 2.4 to 8.7% over the same period . Carcinoid tumors of the small intestine, including duodenum, show similar increases: small intestine is now the major site of gastrointestinal carcinoids.
These increases in incidence rates have been observed in several large analyses in western countries. Beginning slowly in the era of H2RA therapy, the greatest increases have paralleled the widespread marketing of PPIs, a trend observed by many experts [32,33]. Unfortunately, the databases from which these observations were drawn did not record the use of medications, precluding case–control studies. Although the statement has often been made that there have been no cases of gastric carcinoids in humans using PPIs, this is not strictly correct : what can be said is that the available reports in all cases lack information essential to confidently assessing their merit. Beyond this, with population carcinoid incidences in the order of one to two per 1000 people (autopsy rate 1%), prospective studies would require huge numbers of individuals in order to perform statistical analyses with power adequate to demonstrate meaningful differences in their incidence rates. The fraction of PPI users who might demonstrate such a change is unknown. However, with the contemporary growth in large data banks, that now record medication use as well as clinical information, well designed case–control studies could prove of great value in exploring this important issue. The scientific basis for expecting long-term PPI use to cause carcinoid tumors is quite strong and merits serious attention. Hypergastrinemia may also stimulate carcinoid development or growth in other sites. At a minimum, it seems reasonable to discontinue PPI therapy in patients with carcinoid tumors.
Food allergies and eosinophilic esophagitis
Experiments show that a number of ingested potential food allergens that are normally degraded by acid-peptic digestion, become antigenic when the gastric pH rises to values seen during PPI therapy . Other investigations have shown that PPIs dose dependently increase mucosal permeability to small molecules that could include peptide antigens . Adults treated with PPIs for 3 months develop a rise in plasma IgE levels, new food-specific IgE antibodies and a mucosal immune response to offending allergens. A recent hypothesis generating much interest proposes that PPIs may play a key etiologic role in causing eosinophilic esophagitis, as both PPI use and the incidence of eosinophilic esophagitis have risen over closely similar periods .
Effects on other medications
Rises in pH and lack of gastric acid have been shown to impair the absorption of various drugs in common use, including ketoconazole, indinavir, midazolam, didanosine and methadone. Following absorption, some but not all PPIs inhibit various components of the cytochrome P-450 (CYP) enzyme system in liver and intestine, notably CYP 2C19 and CYP 3A4. This has varying effects, most of them subclinical, on many other drugs in common use, but the effects have not risen to the level of posing clinically significant problems. PPIs also variably affect the hepatic catabolism of thyroxine by UDP-glucuronyl transferase, increasing TSH slightly and in some cases requiring an increase in the dose of thyroxine . Another interaction, currently causing a flurry of attention because of its possible linkage to adverse clinical outcomes, is the effect of PPI therapy in patients also using clopidogrel.
Clopidogrel is a prodrug used to inhibit platelet aggregation in patients at increased cardiovascular risk. After absorption, it requires activation by CYP 2C19, before becoming active against platelets. Some PPIs, notably omeprazole and esomeprazole, are metabolized to various extents by CYP 2C19 and this diminishes the activation of clopidogrel, but this is not believed to be a class effect of all PPIs. What is at the heart of current controversy is whether or not this effect on platelet activation translates into increased risks of myocardial infarction, arterial stent occlusion or other related thrombotic risks including death. There have been several large studies, all potentially flawed in various ways, which have yielded conflicting and contradictory conclusions. Even among those observational studies that claim that risk is increased, the ORs or hazard ratios are generally low (under 2.0) and not enough to confidently exclude residual confounding by comorbidity, cotherapy, population genetic heterogeneity and other factors. The studies are especially problematic in their lack of accounting for genetic polymorphisms in CYP 2C19. These affect the metabolism of clopidogrel, contributing to clopidogrel resistance in approximately 30% of the patients [36•]. There are also 21 polymorphisms of CYP 2C19, giving rise to the existence of two main PPI patient groups, extensive metabolizers and poor metabolizers, variously responding to different PPIs [37,38•]. The frequencies of occurrence of such polymorphisms are also different in different races and geographical areas, but, so far, not addressed in resolving problems with clopidogrel activation.
Cotherapy with clopidogrel and low-dose PPI therapy is widely used to minimize the risk of serious gastrointestinal bleeding, particularly in high-risk patients, so a balancing of risks in the individual patient is appropriate. Although the FDA has recently promulgated some cautionary statements , these remain controversial, and many feel that the risk of cotherapy with low-dose omeprazole (20 mg per day) may be acceptable, especially in those at high risk of gastrointestinal bleeding, and particularly if the daily doses are given 12 h apart. The agency's acknowledgement that H2RAs like famotidine, ranitidine or nizatidine (but not cimetidine) can be used instead of PPIs in cotherapy until the issue is resolved also seems prudent. Several studies have shown reductions in hospitalization for gastrointestinal bleeding in patients taking both clopidogrel and PPI compared with those taking clopidogrel alone (COGENT trial results at http://www.theheart.org/article/1007145.do) . With pending emergence of newer antiplatelet agents that do not require activation by CYP 2C19 or interact with PPIs, the whole issue may prove short-lived.
PPIs enjoy an enviable record of safety, but like many drugs that experience a huge increase, particularly in long-term users, after initial marketing, subtle side-effects with low incidences but considerable clinical importance may gradually surface. These ADRs are often relevant only to small numbers of cases or cause only minor problems, but nevertheless they demand widespread education of providers as to their occurrence and management. The true importance of these kinds of events can only be estimated from careful prospective and, when possible, randomized, studies designed solely to measure safety, with minimal confounding. The effects of PPIs now demand such studies, especially addressing the ADRs already associated with their use.
References and recommended reading
Papers of particular interest, published within the annual period of review, have been highlighted as:
• of special interest
•• of outstanding interest
Additional references related to this topic can also be found in the Current World Literature section in this issue (pp. 671–672).
1 Raghunath AS, O'Morain C, McLaughlin RC. The long-term use of proton pump inhibitors. Aliment Pharmacol Ther 2005; 22(Suppl 1):55–63.
2• Ali T, Roberts DN, Tierney WM. Long-term safety concerns with proton pump inhibitors. Am J Med 2009; 122:896–903. An important source of references.
3• Lodato F, Azzaroli F, Turco L, Mazella N, et al
. Adverse effects of proton pump inhibitors. Best Prac Res Clin Gastroenterol 2010; 24:193–201. An important source of references.
4• Thompson ABR, Sauve MD, Kassam N, Kamitakahara H. Safety of the long-term use of proton pump inhibitors. World J Gastroenterol 2010; 16:2323–2330. An important source of references.
5 Yang YX, Lewis JD, Epstein S, Metz DC, et al
. Long-term proton pump inhibitor therapy and the risk of hip fracture. JAMA 2006; 296:2947–2953.
6 Vestergaard P, Rejnmark L, Mosekilde L. Proton pump inhibitors, histamine H2 receptor antagonists, and other antacid medications and the risk of fracture. Calcif Tissue Int 2006; 79:76–83.
7 Targownik LE, Lix LM, Metge CJ, Prior HJ, et al
. Use of proton pump inhibitors and risk of osteoporotic fractures
. CMAJ 2008; 179:319–326.
8 Gray SL, LaCroix AZ, Larso J, Robbins J, et al
. Proton pump inhibitor use, hip fracture, and change in bone mineral density in postmenopausal women. Arch Int Med 2010; 170:765–771.
9• Targownik LE, Lix LM, Leung S, Leslie WD. Proton pump inhibitor use is not associated with osteoporosis or accelerated bone mineral density loss. Gastroenterology 2010; 138:896–904. The best examination of role of bone density in PPI-associated fractures
10 Merriman NA, Putt ME, Metz DC, Yang YX. Hip fracture risk in patients with a diagnosis of Pernicious Anemia. Gastroenterology 2010; 138:1330–1337.
11 Kaye JA, Jick H. Proton pump inhibitor use and risk of hip fractures
in patients without major risk factors. Pharmacotherapy 2008; 28:951–959.
12•• Corley DA, Kubo A, Zhao W, Quesenberry C. Proton pump inhibitors and Histamine-2 receptor antagonists are associated with hip fractures
among at-risk patients. Gastroenterology 2010; 139:93–101. Very likely a key insight into why PPIs may cause fracture in some but not all users.
13 Larner AJ, Hamilton MIR. Review article: infective complications of therapeutic gastric acid inhibition. Aliment Pharmacol Ther 1994; 8:579–584.
14•• Dial MS. Proton pump inhibitor use and enteric infections
. Am J Gastroenterol 2009; 104(S2):S10–S16. The best review of enteric infections
and infestations in patients with hypochlorhydria.
15 Sonnenberg A. Temporal trends and geographical variations of peptic ulcer disease. Aliment Pharmacol Ther 1995; 9(Suppl 2):3–12.
16 Leonard J, Marshall JK, Moayyedi P. Systematic review of the risk of enteric infection in patients taking acid suppression. Am J Gastroenterol 2007; 102:2047–2056.
17 Howell MD, Novack V, Grgurich P, Souillard D, et al
. Iatrogenic gastric acid suppression and the risk of nosocomial Clostridium difficile
infection. Arch Int Med 2010; 170:784–790.
18 Linsky A, Gupta K, Lawler E, Fonda J, et al
. Proton pump inhibitors and risk for recurrent Clostridium difficile
infection. Arch Int Med 2010; 170:772–778.
19 Sultan N, Nazareno J, Gregor J. Association between proton pump inhibitors and respiratory infections
: a systematic review and meta-analysis of clinical trials. Can J Gastroenterol 2008; 22:761–766.
20 Murray LJ, Gabello M, Rudolph DS, Farrell CP, et al
. Transmucosal gastric leak induced by proton pump inhibitors. Dig Dis Sci 2009; 54:1408–1417.
21 Lombardo L, Foti M, Ruggia O, Chiecchio A. Increased incidence of small intestinal bacterial overgrowth during proton pump inhibitor therapy. Clin Gastoenterol Hepatol 2010; 8:504–508.
22 Bajaj JS, Zadvornova Y, Heuman DM, Hafeezullah M, et al
. Association of proton pump inhibitor therapy with spontaneous bacterial peritonitis in cirrhotic patients with ascites. Am J Gastroenterol 2009; 104:1130–1134.
23 McColl KEL. Effect of proton pump inhibitors on vitamins and iron. Am J Gastroenterol 2009; 104:S3–S9.
24•• Waldum HL, Qvigstad G, Fossmark R, et al
. Rebound hypersecretion of acid from a physiological, pathophysiological and clinical point of view. Scand J Gastroenterol 2010; 45:389–394. A critical highly informed review of potential clinical consequences of hypergastrinemia
25•• Reimer C, Sondergard B, Hilsted L, Bytzer P. Proton-pump inhibitor therapy induces acid related symptoms in healthy volunteers after withdrawal of therapy. Gastroenterology 2009; 137:80–87. A demonstration that PPI use causes rebound hypersecretion of acid after therapy stops.
26• Sanaka M, Yamamoto T, Kuyama Y. Effects of proton pump inhibitors on gastric emptying: a systematic review. Dig Dis Sci 2010; 55:2431–2440. A review of effects of PPI therapy on gastric emptying of solids and clinical implications.
27 Kuipers EJ. Proton pump inhibitors and gastric neoplasia. Gut 2006; 55:1217–1221.
28 Kuipers EJ. Barrett's esophagus, proton pump inhibitors and gastrin: the fog is clearing. Gut 2010; 59:148–149.
29 Wang JS, Varro A, Lightdale CJ, Lerkowit N, et al
. Elevated serum gastrin is associated with a history of advanced neoplasia in Barrett's esophagus. Am J Gastroenterol 2010; 105:1039–1045.
30 Modlin IM, Lye KD, Kidd M. A 5-decade analysis of 13,715 carcinoid
tumors. Cancer 2003; 97:934–959.
31 Modlin IM, Lye KD, Kidd M. A 50-year analysis of 562 gastric carcinoids: small tumor or larger problem? Am J Gastroenterol 2004; 99:23–32.
32 Hodgson N, Koniaris LG, Livingstone AS, Franceschi D. Gastric carcinoids: a temporal increase with proton pump inhibitor introduction. Surg Endosc 2005; 19:1610–1612.
33 Waldum H, Gustafsson B, Fossmark R, Qvigstad G. Antiulcer drugs and gastric cancer. Dig Dis Sci 2005; 50:S39–S44.
34 Merwat SN, Spechler SJ. Might the use of acid-suppressive medications predispose to the development of eosinophilic esophagitis? Am J Gastroenterol 2009; 104:1897–1902.
35 Sachmechu I, Reich DM, Wibowo AM, et al
. Effect of proton pump inhibitors on serum thyroid-stimulating hormone level in euthyroid patients treated with levothyroxine for hypothyroidism. Endocr Pract 2007; 13:345–349.
36• Mega JL, Close SL, Wiviott SD, Shen L, et al
. Cytochrome P-450 polymorphisms and response to clopidogrel. N Engl J Med 2009; 360:354–362. A good discussion of the relevance of genetic polymorphisms in resistance to clopidogrel.
37 Chaudry AS, Kocchar R, Kohli KK. Genetic polymorphism of CYP2C19 and therapeutic response to proton pump inhibitors. Ind J Med Res 2008; 127:521–530.
38• Klotz U. Impact of CYP2C19 polymorphisms on the clinical action of proton pump inhibitors (PPIs). Euro J Clin Pharmacol 2009; 65:1–2. A good discussion of the relevance of genetic polymorphisms to the clinical effects of PPIs.
39 Alkhatib AA, Elkhatib FA, Khatib OF. Gastric acid-reducing medications and clopidogrel: what are the latest FDA recommendations? Am J Gastroenterol 2010; 105:1211.
40 Ray WA, Murray KT, Griffin MR, Chung CP, et al
. Outcomes with concurrent use of clopidogrel and proton pump inhibitors. Ann Int Med 2010; 152:337–345.