Long-term Use of Proton-Pump Inhibitor Therapy

G&H Who are the patients that generally require maintenance dosing of proton-pump inhibitors (PPIs)?

DM The two main indications for long-term use of PPIs are reflux disease and use of maintenance nonsteroidal anti-inflammatory drugs (NSAIDs), which puts patients at risk for nonsteroidal gastropathy. Except for hypersecretory states, which are very rare disorders, most other indications for acid suppression do not require years and years of PPI exposure.

For patients who are taking PPIs long-term, the likelihood of requiring continued therapy depends to some extent on the size of the initial dose. Patients requiring higher initial doses to achieve effective symptom relief are more likely to require continued therapy as well. This is in part because higher dosing implies a more severe abnormality requiring greater acid suppression for symptom relief (eg, gastroesophageal reflux disease in the setting of erosive esophagitis with a poor gastroesophageal barrier). The treatment itself may also predispose patients to a need for ongoing therapy. In suppressing acid, PPIs stimulate the body's feedback loop that tries to reactivate acid secretion. If the drug is removed, there is a potential risk of rebound hypersecretion, creating a sort of dependency on the drug because the body is acclimated to having acid suppressed. In addition, ongoing feedback stimulation creates a need for ongoing therapy to control symptoms, and higher initial doses are more likely to activate this feedback response. This is one reason why it is important to start patients, when clinically indicated, on the lowest effective maintenance dose of PPI. However, these drugs do not engender tachyphylaxis, or loss of efficacy, over time, and patients rarely require dose escalation to maintain efficacy.

G&H What guidance does the current approved prescribing information literature for PPIs provide regarding long-term use of these drugs?

DM None of the prescribing information verbiage recommends a specific time span for maintenance therapy. The official labels for lansoprazole (Prevacid, TAP), pantoprazole (Protonix, Wyeth), rabeprazole (Aciphex, Ortho McNeil), and omeprazole/sodium bicarbonate (Zegerid, Santarus) state that “controlled studies do/did not extend beyond 12 months.” For esomeprazole (Nexium, AstraZeneca), the label states that “controlled studies do not extend beyond 6 months.” The drug facts panel for omeprazole (Prilosec OTC, Procter & Gamble), an over-the-counter medication, states that “consumers are advised to take no more than one 14-day course of treatment of therapy every 4 months unless otherwise directed by a physician.” However, many studies of each of these agents have looked at longer-term data. The US Food and Drug Administration (FDA) has requested that manufacturers supply 3-year data if available, and most of the makers of PPIs are complying. In addition, investigators around the world have published many studies looking at longer-term exposure beyond labeled indications.

G&H Can you describe the potential adverse effects of long-term PPI therapy?

DM Idiosyncratic reactions occur, as with any drug, on very rare occasions and are unpredictable. Because of the way PPIs are metabolized, genetic polymorphisms (ie, normal variants among occasional patients) may increase the risk of hepatotoxicity, nephrotoxicity, or other unusual side effects. These sorts of effects are thankfully very rare. More common, however, are potential adverse effects relating to the two known class effects of PPIs: hypochlorhydria and hypergastrinemia. PPIs are designed to suppress acid secretion, which leads to the question of whether a complete absence of acid (achlorhydria) or a reduction in acid production (hypochlorhydria) can be potentially harmful. Also, the feedback loop wherein acid suppression signals the body to attempt to restore gastric acid production through hypergastrinemia, which is an increase in the blood level of the hormone gastrin, may also predispose to long-term effects.

G&H What are the possible hypochlorhydria-related effects of PPIs?

DM The phenomenon of hypochlorhydria has been linked to an increased susceptibility to bacterial or parasitic infections. However, several clinical scenarios where this might have an effect in relation to PPI use have been investigated and have not proven clinically relevant. Patients on PPIs who travel to developing countries have been observed to see if they have a higher rate of travelers' diarrhea, but any raised incidence has not been shown to be significant. We know that gastric acid increases the size of the inoculum required to contract typhoid, for example. In the absence of the gastric acid barrier, a lower inoculum may still cause clinical infection. Again, this observation has not proven to be clinically relevant in relation to PPI use in general.

The current in-hospital epidemic of Clostridium difficile-related colitis has been observed in large epidemiologic studies to have an association with PPI use, although the odds ratios for this association tend to be much lower than those observed for other risk factors such as antibiotic exposure. Does this mean that patients should stop their PPIs if they are admitted to the hospital? By doing so, patients become exposed to the risk of rebound hypersecretion and recurrence of heartburn or bleeding from ulcers. Thus, some balance of risk and benefit must be sought. The major risk factor for C. difficile infection in the hospital setting is poor handwashing techniques that allow for the passage of spores from person to person. In my opinion, proper barrier nursing is more important than PPI use in controlling this problem. It should be noted that a small but statistically significant association between PPI use and community-acquired C. difficile infection has also been described, supporting my general recommendation for using the lowest effective maintenance dose of PPIs when clinically indicated.

In addition, several studies of acid suppression support the theory that aspirated gastric contents (colonized in the absence of acid) may cause an increased rate of ambulatory pneumonia in long-term PPI users. Again, these studies have yielded very small odds ratios that are most likely not clinically important, but this is another issue to consider.

Our group conducted a study that identified an association between long-term PPI use and hip fracture. This was, again, a small but statistically significant finding supported by the presence of a dose response (ie, longer or more potent exposure to acid suppression had a higher correlation with hip fracture). We think this is probably due to hypochlorhydria altering the systemic absorption of calcium. All people, as they become older, excrete more calcium in the urine, absorbing less systemically. Adding a PPI therapy may magnify this effect. The same is potentially true of vitamin B₁₂ absorption. In the presence of drug-induced hypochlorhydria, it has been theorized that patients may not make enough intrinsic factor to absorb B₁₂ appropriately. Thus, elderly people on PPIs are potentially at risk for both B₁₂ deficiency affecting their steadiness, which in turn predisposes them to falling and breaking their hips from drug-associated osteoporosis. For both of these scenarios, the same mantra holds true: do not deny PPIs to those who need them, but use the lowest effective maintenance dose. More important for all elderly patients, whether on or off PPI therapy, is the need to monitor bone density and make sure they are taking supplemental calcium and B₁₂, if necessary.

G&H What are the effects that could be attributed to PPI-related hypergastrinemia?

DM With regard to hypergastrinemia, elevated gastrin levels are associated, in theory, with increased turnover of cells and possibly a predisposition to cancer, though this has never been demonstrated to be clinically relevant. Even in prototypical conditions associated with very high gastrin levels, such as Zollinger-Ellison syndrome or pernicious anemia, hypergastrinemia per se has not been shown to cause clinically relevant tumors. Early studies of PPI therapy in animal models (primarily rats) demonstrated a propensity for gastric carcinoid formation, presumably from drug-induced hypergastrinemic stimulation of enterochromaffin-like cells in the gastric mucosa. This has not been shown to occur in humans unless otherwise predisposed (eg, MEN-1 syndrome or pernicious anemia).

In the past, researchers have tentatively linked high gastrin levels to esophageal, pancreatic, colon, and lung cancer. None of these theories has borne fruit after many years of exposure, although the potential association with gastric cancer is of interest. The longest follow-up study published to date (studying up to 12 years of exposure) was conducted by Dr. Klinkenberg-Knol and associates in the Netherlands. None of their patients developed any problems related specifically to elevated gastrin levels as a result of PPI therapy. However, there have been suggestions, in other studies of this cohort following patients for approximately 7 years, that PPI therapy in patients with Helicobacter pylori infection may together potentially cause atrophy of the stomach. Atrophy could then predispose patients to dysplasia and, ultimately, gastric cancer. It should be stressed that the current data on this phenomenon are very limited and general guidelines do not call for testing for H. pylori before starting patients on PPIs. Currently, I consider this a theoretical concern that requires further study before any treatment recommendations can be made.

Another potential association of PPI therapy with hypergastrinemic side effects is fundic gland polyps, which are benign polyps in the stomach. Very occasionally, these polyps are of concern if they are associated with a hereditary colorectal cancer condition called familial adenomatous polyposis. However, this hereditary condition is extremely uncommon and unrelated to gastroesophageal reflux. There is some anecdotal evidence that fundic gland polyps may be increasing in incidence in patients taking PPIs. However, it is currently unclear as to whether an increase in this generally benign condition will result in any clinically relevant adverse effects.

G&H Could you describe the idiosyncratic and metabolic adverse effects that have been documented with PPI use?

DM Studies from Australia have noted isolated incidents of interstitial nephritis related to PPIs. This is the type of infrequent reaction that can be associated with any drug and is not related to the duration of use. Liver disease is another rare associated condition. Anaphylaxis and allergy have likewise been described on rare occasions.

Cytochrome P450 interactions can lead to potential toxicity with other coprescribed drugs (such as warfarin, theophylline, and diazepam), but blood levels can be measured at the lowest effective maintenance PPI dose and need not cause concern if they are normal. More recently, a paper was published looking at the potential P450-mediated competitive interaction of PPIs with clopidogrel (Plavix, BMS/Sanofi Aventis), a blood thinner used to promote stent patency in cardiovascular patients. The authors noted that omeprazole taken with clopidogrel can reduce clopidogrel efficacy. Again, this is a preliminary study in the in vitro setting, which may not have clinical relevance, but it requires further research.

G&H In a patient who requires dose escalation of PPIs over time, do concerns of adverse events warrant the use of alternative (ie, surgical) therapy?

DM First of all, if PPIs are effective, the dose should not require escalation over time because there is no tachyphylaxis associated with these drugs (in fact, it may even be possible to reduce maintenance doses with time). Such patients may consider surgery to reduce their drug costs or because they choose not to take therapy, but they should not be considered PPI failures. Those who require dose escalation for symptom control should be tested with pH monitoring to ensure that they truly have failed PPI therapy, as they could have some other condition beyond gastroesophageal reflux at the root of their symptoms. Patients with true reflux who are PPI failures are extremely uncommon, though this scenario could conceivably represent a reasonable indication for surgical fundoplication; these patients have essentially failed medical management of their reflux disease. However, individual patients benefit from fundoplication to varying extents. If the goal is to cease medical therapy because patients have been taking PPIs for too long at too high a dose, what guarantee does surgical therapy provide that patients can remain off PPIs after they have undergone surgery? Overall, I am very wary of recommending fundoplication in patients who do not respond to PPIs because of the concern that they may not have reflux disease at all. There is a very small group of patients who have weakly acidic or nonacidic reflux. These patients are candidates for surgery, but that decision requires further consultation with an expert in the field.

It should, however, be noted that all PPIs have FDA-approved dosage levels for maintenance. In many cases, the approved maintenance dose is half of the full dose, taken once a day. Although we know that approximately 15–20% of patients take less than a full dose or use PPIs in an on-demand fashion, another 10–15% require offlabel double dosing to maintain symptom remission. Patients on an elevated off-label dose should be warned that the higher the dose, the more likely they are to have hypergastrinemia, hypochlorhydria, and any potential attendant side effects.

G&H Which, if any, of the above side effects require further study in long-term PPI use?

DM The osteoporosis issue is potentially important because patients who fracture their hips have a 20% chance of death. However, if the osteoporosis concern can be overcome by taking maintenance calcium, it could be negated; this is the scenario that requires further study.

Interaction with clopidogrel is also potentially important because all the antiplatelet therapies utilized in cardiovascular medicine can result in upper gastrointestinal bleeding and ulcers, which potentially require PPI therapy. If our therapy to prevent bleeding makes antiplatelet therapy less effective, some compromise must be reached.

The H. pylori/PPI interaction in gastric cancer could potentially become important in countries with widespread H. pylori infection and a rising incidence of reflux such as India, China, and Malaysia. In the United States, H. pylori has been nearly eradicated and is not of much concern.

The most important point to make is that these drugs have an overall superb efficacy and safety profile, and they should not be denied to patients who will benefit from their use. All of the concerns outlined above are theoretical and though they may require further research, they should not affect current clinical practice.