Taking a PPI for years: vitamin B12, magnesium and long-term side effects

The best-documented signals concern the absorption of certain nutrients, vitamin B12, magnesium and iron, as well as a higher risk of fracture^1,2,3,4. Other associations reported by observational studies were not found in a large randomised trial⁵.

This trial, called COMPASS, randomly assigned 17,598 patients with stable cardiovascular disease to 40 mg of pantoprazole a day or a placebo. They were followed for a median of 3.01 years. The researchers tracked pneumonia, fractures, dementia, cancers, chronic kidney disease, diabetes and mortality. They found no significant difference, except for intestinal infections: 1.4% on pantoprazole versus 1.0% on placebo⁵.

This trial provides the strongest data available, but it has its limits. Three years is shorter than many real-world treatments, and the list of outcomes tracked did not include blood levels of vitamin B12 or magnesium⁵. The nutritional signals described below are therefore neither confirmed nor ruled out by this trial.

In food, vitamin B12 is bound to proteins, and stomach acid helps release it. By sharply reducing this acidity, a PPI can lower the absorption of B12 from the diet^6,7.

Absorption takes place in several steps. In the stomach, hydrochloric acid and a digestive enzyme detach B12 from its food matrix. It then binds to a carrier protein, then to intrinsic factor secreted by the stomach, and this complex is absorbed at the end of the small intestine⁶. The omeprazole leaflet states that lower acidity may reduce vitamin B12 absorption, and that this should be taken into account in people with low stores or other risk factors⁷.

The vitamin B12 added to fortified foods and supplements is already in free form and does not need this release step. People who absorb food B12 poorly absorb free B12 normally, according to the US National Institutes of Health⁶.

This study found that taking a PPI for two years or more was associated with a 1.65 times higher risk of vitamin B12 deficiency in more than 25,000 deficient patients¹.

Researchers at Kaiser Permanente in Northern California compared 25,956 patients diagnosed with B12 deficiency between 1997 and 2011 with 184,199 patients without deficiency. Among the deficient group, 12.0% had received at least two years of PPIs, compared with 7.2% of the others. The odds ratio reached 1.65 (95% confidence interval: 1.58 to 1.73). H2 receptor antagonists, which reduce acidity less, were associated with a lower risk, with an odds ratio of 1.25¹.

A dose effect also emerged. Above 1.5 tablets a day, the odds ratio rose to 1.95, compared with 1.63 below 0.75 tablets a day¹. This is a case-control study: it shows a strong and consistent association, without establishing that a PPI causes the deficiency in any given person.

Signs of B12 deficiency develop slowly and can take several years to appear. They include fatigue, neurological problems such as tingling, palpitations, an inflamed tongue and anaemia with enlarged red blood cells⁶.

Lower levels have been measured for both, with levels of evidence comparable to those for B12: numerous observational studies, without a controlled trial on these outcomes^2,3,8.

For magnesium, the FDA added a warning to all prescription PPIs in 2011. In most reported cases, hypomagnesaemia had appeared after more than a year of treatment². A meta-analysis of sixteen studies and 131,507 patients found hypomagnesaemia in 19.4% of users compared with 13.5% of non-users, with an adjusted odds ratio of 1.71⁸. Agencies suggest considering a magnesium test before prolonged treatment, then regularly, especially when a diuretic or digoxin is also taken^2,7.

For iron, the same Kaiser Permanente team compared, in 2017, 77,046 patients with iron deficiency with 389,314 controls, excluding people already exposed to a known risk factor. Two or more years of PPIs were associated with an adjusted odds ratio of 2.49 (2.35 to 2.64). The association increased with dose and decreased after the medicine was stopped. It reached 4.27 in people who had taken more than 1.5 tablets a day for at least ten years³.

Meta-analyses of observational studies link PPI use to a moderately higher fracture risk, in the order of 20 to 30% for the hip^4,9. The COMPASS trial found no difference over three years⁵.

In 2010, the FDA changed the leaflets of all PPIs to mention a possible risk of hip, wrist and spine fractures. In the studies reviewed, the highest risk concerned high doses or treatments of one year or more, especially in people aged 50 and over¹⁰.

A 2016 meta-analysis of eighteen studies and 244,109 fractures found a relative risk of 1.26 for the hip, 1.58 for the spine and 1.33 for all fractures combined. Heterogeneity between studies remained notable, and the authors observed no effect of duration: less than one year and more than one year of treatment gave similar risks⁴. In 2019, another meta-analysis of 24 studies and more than two million participants found a relative risk of 1.20 for the hip, more marked at high doses⁹.

The link with calcium remains a hypothesis. The absence of a duration effect fits poorly with a simple deficiency building up over the years, and confounding factors such as age, frailty or other treatments may weigh on these studies. The risk deserves to be taken into account in someone whose bones are already fragile, without being regarded as proven.

The French National Authority for Health (HAS) asks that every PPI prescription be reviewed: effectiveness, quality of life, adverse effects and drug interactions¹¹. It estimates that more than half of uses in France are not justified.

For uncomplicated reflux, the HAS sets an initial treatment of four weeks and considers continuing it rarely justified, particularly in older people taking several medicines. Before extending beyond three months, it recommends investigating the cause and looking for the lowest effective dose. It also notes that rebound acidity is possible on stopping, without justifying a new prescription on its own¹¹.

The American Gastroenterological Association also recommends a regular review of the indication, documented by the treating doctor. It specifies that some people should generally not stop their PPI, for example in the case of Barrett’s oesophagus, severe oesophagitis or a high risk of gastrointestinal bleeding. It adds that the decision to stop should be based on the absence of an indication, not on fear of an adverse effect¹². Stopping or changing a PPI is therefore always decided with the doctor or pharmacist.

It is worth telling the doctor or pharmacist about the duration of treatment, the dose, other medicines taken and any new symptom. This information guides the review and any blood tests.

• The duration and dose of the PPI, including packs bought without a prescription, which the FDA asks you to mention².
• Taking a diuretic or digoxin, which makes a magnesium test worth considering^2,7.
• Persistent fatigue, cramps, tingling, palpitations or tremors, which can have many causes^2,6.
• Any known situation of low B12 stores or poor absorption, for which the omeprazole leaflet calls for particular attention⁷.
• A fracture, known osteoporosis or plans to stop treatment^10,11.

Whether to test vitamin B12, magnesium or iron is decided on the basis of these factors. Adding a supplement on top of the treatment is discussed in the same way, taking into account the results and other medicines.