What are the potential concerns with long-term use of Proton Pump Inhibitors (PPIs)?

Potential Concerns About Long-Term PPI Use

While observational studies have suggested associations between long-term PPI use and various adverse outcomes including infections, fractures, and micronutrient deficiencies, no randomized controlled trial has demonstrated that PPIs increase the rate of any adverse event, and these concerns should not be used as justification to discontinue PPIs when a valid indication exists. 1

Evidence Quality and Clinical Context

The critical distinction in evaluating PPI safety concerns is the quality of evidence:

• All studies reporting specific associations between PPIs and serious adverse events (chronic kidney disease, fractures, dementia, COVID-19) have been observational and cannot establish causality 1
• Randomized controlled trials comparing PPIs with placebo have consistently shown no higher rate of adverse events among PPI users 1
• Many reported associations lack plausible mechanisms of action and are likely explained by residual confounding and analytic biases 1

Established Adverse Effects with Higher Probability of Causality

Gastrointestinal Infections

• Increased risk of Clostridium difficile-associated diarrhea, particularly in hospitalized patients 2
• Increased risk of community-acquired pneumonia (67% higher odds), though not hospital-acquired pneumonia 3
• PPIs may increase susceptibility to gastroenteritis and other enteric infections due to reduced gastric acid barrier 1

Rebound Acid Hypersecretion

• Common after discontinuation of long-term PPI therapy, lasting 2-6 months, representing a physiological response to secondary hypergastrinemia 1, 4
• This rebound effect may necessitate continued PPI use for symptom suppression 5

Acute Tubulointerstitial Nephritis

• Can occur at any point during PPI therapy, presenting with varying signs from hypersensitivity reactions to non-specific decreased renal function 2
• Requires discontinuation and evaluation if suspected 2

Associations with Weaker or Conflicting Evidence

Bone Fractures

• Meta-analysis of observational studies showed 42% increased risk of hip fracture (OR=1.42; 95% CI: 1.33-1.53) and 20% increased risk overall (RR=1.20; 95% CI: 1.14,1.28) 6, 3
• However, large RCTs including the COMPASS trial found no differences in fracture rates between PPI and placebo groups 6
• The association appears strongest in patients with pre-existing risk factors (diabetes, CKD, arthritis) and ≥2 years of use 6
• FDA includes precautionary notices regarding fracture risk with high-dose, long-term use (≥1 year) 2

Micronutrient Deficiencies

Vitamin B12:

• Large RCTs have not shown significant differences in serum B12 levels at 5 years, though these studies had methodological limitations 6
• FDA warns that daily treatment >3 years may lead to malabsorption due to hypo- or achlorhydria 2
• Deficiency appears rare except possibly in elderly patients or those with Zollinger-Ellison Syndrome on high doses 5

Magnesium:

• Meta-analysis shows 71% higher risk of hypomagnesemia (adjusted OR: 1.71; 95% CI: 1.33,2.19) 6
• FDA includes precautionary notices regarding anemia risk 6

Iron:

• Dose-dependent associations exist between continuous PPI use and iron deficiency, particularly after ≥1 year of use 6
• Reduced gastric acid impairs absorption of non-heme iron 6

Other Reported Associations

Cutaneous and Systemic Lupus Erythematosus:

• Both CLE and SLE reported in patients taking PPIs, occurring as new onset or exacerbation of existing disease 2
• Most common form is subacute CLE (SCLE), occurring within weeks to years 2
• Most patients improve with discontinuation alone in 4-12 weeks 2

Cardiovascular Risk:

• Some observational studies suggest increased cardiovascular disease risk with long-term use 6
• Avoid concomitant use with clopidogrel due to CYP2C19 inhibition reducing clopidogrel's antiplatelet activity 2

Cancer Risk:

• Observational studies suggest 55% higher odds of colorectal cancer (OR=1.55; 95% CI: 0.88-2.73) 3
• Japanese population-based data suggest possible association with gastric cancer, though rates similar to PPIs versus H2-receptor antagonists 1
• No causal relationship established in RCTs 1

Enterochromaffin-Like Cell Hyperplasia:

• Demonstrated in up to 50% of children receiving PPIs for >2.5 years, though considered a benign histologic change 1
• Five-year RCT comparing vonoprazan and lansoprazole found infrequent and comparable proportions developing ECL hyperplasia 1

Critical Management Principles

When NOT to Discontinue PPIs

Patients with the following definitive indications should continue long-term PPI therapy: 1, 7, 6

• Barrett's esophagus
• Severe erosive esophagitis (LA Classification grade C/D)
• History of esophageal ulcer or peptic stricture
• Gastroprotection in high-risk NSAID/aspirin users
• Secondary prevention of gastric/duodenal ulcers
• High risk for gastrointestinal bleeding
• Zollinger-Ellison Syndrome

When to Consider De-prescribing

• All patients without a definitive indication for chronic PPI should be considered for trial of de-prescribing 1, 7
• Most patients on twice-daily dosing should be stepped down to once-daily PPI 1
• Regular review of ongoing indications should be performed by the primary care provider 1, 7

Key Clinical Pitfall

Discontinuing PPIs in patients with definite indications based on concerns about unproven risks may lead to recurrent symptoms and serious complications, including upper GI bleeding 1, 4

Practical Approach to Safety Monitoring

For patients requiring long-term PPI therapy:

• Use the lowest effective dose and shortest duration appropriate for the condition 2
• Document the ongoing indication clearly 1, 7
• Monitor for signs of acute TIN (decreased urine output, blood in urine) 2
• Consider monitoring magnesium levels in high-risk patients 6
• Evaluate for C. difficile if diarrhea develops that does not improve 2
• Assess for lupus symptoms (new joint pain, photosensitive rash) 2
• Do not discontinue PPIs solely due to concern about potential adverse events when a valid indication exists 1, 4