Reported Side Effects and Complications of Long-term Proton Pump Inhibitor Use

Dissecting the Evidence

David A. Johnson; Edward C. Oldfield IV


Clin Gastroenterol Hepatol. 2013;11(5):458-464. 

In This Article

Proton Pump Inhibitors and Infection


Several studies have focused on assessing the risk between PPI use and community-acquired pneumonia (CAP) and hospital-acquired pneumonia. The initial case-control study of 5551 cases in The Netherlands found a relative risk for CAP among PPI users of 1.89 (95% CI, 1.32–2.62).[44] Subsequently, 2 other studies found a moderate risk of CAP in patients exposed to PPIs.[45,46] The most recent meta-analysis (9 studies, 120,863 patients) further delineated the relative risks of PPI use and CAP, finding that there was no association between CAP and PPI use longer than 180 days (OR, 1.10; 95% CI, 1.00–1.21); rather, the association between PPI use and CAP was strongest for PPI use of fewer than 30 days (OR, 1.65; 95% CI, 1.25–2.19) and high-dose PPIs (OR, 1.50; 95% CI, 1.33–1.68).[47] Also supporting the association between short-term PPI use and CAP, a database review of 71,985 outpatient prescriptions for PPIs in the New England Veterans Healthcare System found that PPI use between 1 and 15 days had increased risk for CAP over longer PPI exposures.[48]

Despite the results of these studies, other studies have found no significant increase in CAP risk from PPI use, long term or current.[49] A case-controlled review of 80,000 patients found that when accounting for potential confounding factors, there was no significant association between current PPI use and increased CAP (adjusted OR, 1.02; 95% CI, 0.97–1.08).[50] This highlights the influence of heterogeneity between studies and the potential influence of confounding factors on the results of the other studies.

Bottom Line. Health care providers should be aware of the potential adverse relationship between PPI use and CAP, namely, a small relative risk associated with short-term and high-dose PPI use. These relationships, however, do not offer a definitive explanation for the relative risk because significant heterogeneity among studies and a number of confounding factors may have accounted for some of the observed statistical significance.

Clostridium Difficile

Previously, gastric acid was not believed to be important in protecting against C difficile infection because acid-resistant spores were presumed to be the principal vector of transmission. Recently, this thought was challenged because several studies have found a higher risk of C difficile infection in PPI users. In theory, PPIs may increase the risk of C difficile infection by increasing the ability of the spore to convert to the vegetative form and to survive in the lumen of the GI tract. The data for community-acquired vs hospital-acquired infection has been variable and inconclusive for an associated risk of harm.[51] One of the first meta-analysis (11 studies, 127,000 patients) found a significant relationship between PPI use and C difficile infection, with an OR of 2.05 (95% CI, 1.47–2.85).[52] Further supporting the hypothesis of a direct causative association, a recent study found a significant dose response, with more aggressive acid suppression associated with higher ORs.[53] These findings also were supported by another meta-analysis (23 studies, 300,000 patients), which found PPI use was associated with an OR of 1.69 (95% CI, 1.395–1.974).[54] It is important to note that this study had several significant drawbacks including unaccountable heterogeneity and lack of information on potential confounders.

Despite the results of these earlier studies, the most recent studies offered conflicting viewpoints about the association between PPI use and increased risk of C difficile infection. In one study, researchers evaluated the association between acid-suppressing agents (PPIs and H2RAs) in 385 patients who had C difficile infection. Univariate analysis revealed both PPI and H2RA use was associated significantly with increased risk. After adjusting for age and comorbid conditions, however, there was no association with increased incidence or recurrence of C difficile infection.[55] Another case-controlled study in hospitalized patients found that length and dose of PPI exposure was not associated significantly with increased risk of C difficile infection (P = .416); rather, only antibiotic exposure in the past 3 months was associated significantly with C difficile infection (OR, 5.97; 95% CI, 2.40–14.8; P = .001).[56] Of note, the most recent review on detection, prevention, and treatment of C difficile did not include restriction or avoidance of PPIs in the recommendations for prevention of C difficile infection,[57] and this has not been recommended by multisociety clinical practice guidelines.[58]

Bottom Line . To date, there is insufficient evidence to conclude that there is a definitive relationship between PPI use and C difficile infection. Given the increasing prevalence and morbidity associated with this infection, clinicians should be aware of this potential relationship, yet understand that confounding factors may play a significant role in the reported association. Appropriate use of PPIs should not be changed, however, until there is more conclusive evidence for potential harm.

Traveler's Diarrhea

Alterations of the gastric pH and possible related changes in susceptibility for enteric infections have been a topic of long-standing debate. Although gastric hypochlorhydria commonly is listed as a risk factor for traveler's diarrhea,[59] PPI exposure as a risk factor for enteric infections in travelers has not been studied formally. In fact, there is only one study that evaluated acid-reduction medication use and this study reported no significant association (OR, 6.9; range, 0.7–67.4) of traveler's diarrhea with antacids and H2-receptor–antagonist use.[60] A meta-analysis of the diagnosis of enteric infections did identify an increased risk of acute bacterial infection associated with the use of PPIs (OR, 3.33; 95% CI, 1.84–6.02).[52] A recent comprehensive analysis of the data on PPI use and enteric infections concluded that there was no association of PPI use and viral or parasitic enteric infections.[51]

Bottom Line. The data on specific bacterial infections were overall supportive of no associated risk, albeit there were a few specific case reports suggesting a remote causal association. The International Society of Travel Medicine, however, does suggest discontinuing PPIs if traveling to areas with risk of enteric infection.[61] This seems reasonable if patient risk assessment is individualized and, when possible, PPIs can be stopped for a short period of time without other GI consequences.

Small Intestinal Bacterial Overgrowth

Small intestinal bacterial overgrowth (SIBO), a condition that is associated with bloating, diarrhea, and malabsorption, recently has been associated with PPI use, although the significance of the association is uncertain.[53] In this report of 450 patients, SIBO was detected in 50% of patients using PPIs, 24.5% of patients with irritable bowel syndrome (IBS), and 6% of healthy control subjects. There was a statistically significant difference between patients using PPIs and those with IBS or healthy control subjects (P < .001).[62] The prevalence of SIBO increased after 1 year of treatment with PPI. This finding is supported by a smaller study of 42 patients, showing an association within the first 8 weeks of PPI use and also an increasing incidence of SIBO at the 6-month mark (P < .05).[63]

Since that article was published, 2 other retrospective case reviews have suggested no clear association between PPI use and SIBO. The first study, consisting of a database analysis of 675 patients who received a duodenal aspirate, found no clear association between SIBO and either PPI use or IBS (P < .05).[64] In addition, this study reported a positive association between older age (>50 y) and increased incidence of SIBO (OR, 5.7; 95% CI, 3.7–23.5). The second retrospective chart review of 1191 patients also found no association between PPI use and SIBO, using either univariate or multivariate regression.[65] In addition, treatment of SIBO is not impaired significantly in patients with PPI use because the reported eradication rate of SIBO (using rifaximin) was 87% in the PPI group and 91% in the IBS group.[62]

Bottom Line. The relationship between PPI use and the development of SIBO is still not understood. Given the lack of randomized control trial data and reports that have significant confounding bias potentials, there are no clear supporting data at present to suggest a positive relationship.

Spontaneous Bacterial Peritonitis

Recent reports have suggested that there is a relationship between PPI use and the development of spontaneous bacterial peritonitis (SBP) in hospitalized cirrhotic patients with ascites. One study found a strong association (OR, 4.3; 95% CI, 1.3–11.7) between PPIs and SBP,[66] whereas another study found no significant association (OR, 1.0; 95% CI, 0.4–2.6).[67] A recent meta-analysis (4 studies, 772 patients) reported a significant association between PPI use and the development of SBP in cirrhotic patients (OR, 2.77; 95% CI, 1.82–4.23).[68] Given the large sample size compared with other studies on the topic and the low level of heterogeneity (I2 = 22%), the investigators recommended that PPIs should be used judiciously and only when clearly indicated for the cirrhotic patient.

A recent retrospective, propensity-matched cohort study used US Veterans Health Administration data to compare rates of serious infection associated with use of PPIs, H2RAs, or no gastric acid suppressant in patients who began use after the development of decompensated cirrhosis. Serious infections were defined as any infection requiring hospitalization, and a subset of these infections were classified as related to acid suppression (pneumonia, bacteremia, C difficile, and SBP). A total of 4181 patients were included in the analysis (1905 PPI users, 248 H2RA users, and 2028 nonusers of gastric acid suppressants). Compared with nonusers, PPI users had a higher incidence of serious infections (adjusted hazard ratio, 1.66; 95% CI, 1.31–2.12) as well as acid-suppression–related infections (adjusted hazard ratio, 1.75; 95% CI, 1.32–2.34). These results are supportive of the previous findings that PPI use increases the risk for serious infections in patients with decompensated cirrhosis.[69] Recognizably, there is a background risk for patients with cirrhosis, in particular for patients with low protein ascites.[70] These patients have increased relative risks for disruptions in the composition of the GI microflora, owing to medical therapies and abnormal intestinal motility. Evidence suggests that 25% have small-bowel bacterial overgrowth, which can promote intestinal wall permeability that results in bacteria translocation and secondary infections (eg, SBP).

Bottom Line. Although there is no definitive evidence for conclusion, PPI use in the cirrhotic patient should be scrutinized for appropriateness for use. In the studies to date suggesting possible causal harm, the majority of patients did not meet criteria to justify continued use of PPIs. At present, it would be premature to recommend routine discontinuance of PPIs in the patients who have appropriate indications for continued appropriately justified use of PPIs. The most recent data from Bajaj et al[66] suggest that H2RAs do not have this relative risk. Accordingly, if the patient has decompensated cirrhosis and needs continued acid-reduction therapy, it is reasonable to try switching to an H2RA and monitor the clinical effectiveness of the change in therapy.

Interstitial Nephritis

Several case reports have implicated PPIs as a cause of acute interstitial nephritis. This disorder is a humoral and cell-mediated hypersensitivity inflammatory reaction of the renal interstitium and tubules. A systematic review from 2007 found 64 cases documented in the literature, 12 of which were considered certainly associated, and 9 of which were probably associated.[71] Initial symptoms were nonspecific and included nausea, malaise, and fever. With such extensive use worldwide as the denominator, the investigators concluded that acute interstitial nephritis was a rare, idiosyncratic occurrence related to PPI use, but did not find enough evidence to support a causative relationship.

Bottom Line. Despite the extreme rarity of the syndrome, the association cannot be dismissed and a high level of clinical suspicion to detect acute interstitial nephritis early in its course, especially soon after the initiation of PPI therapy, should be followed up.