PPI heartburn drugs 'up hip fracture risk in smokers'

“Heartburn pills taken by thousands of women ‘raise risk of hip fractures by up to 50 per cent’,” the Daily Mail reported today. The headline is based on a large new study of drugs called proton pump inhibitors (PPIs), which are commonly used to treat heartburn, acid reflux and ulcers.

The study found that post-menopausal women who regularly took PPIs for at least two years were 35% more likely to suffer hip fracture than non-users, a figure that increases to 50% for women who were current or former smokers. However, although this increase in risk is large, the overall risk of fractures remains small.

This was a large, well conducted study that suggests that long-term use of PPIs is associated with a small increase in risk of hip fracture, although the researchers point out that the risk seems to be confined to women with a history of smoking. Unlike previous research, this study took careful account of other factors that might affect risk such as body weight and calcium intake.

Women who are concerned about their use of PPIs are advised to consult their GP.

Where did the story come from?

The study was carried out by researchers from Massachusetts General Hospital, Boston University and Harvard Medical School and was funded by the US National Institutes of Health. The study was published in the peer-reviewed British Medical Journal.

Although the Mail’s headline is technically correct, it gives the impression that these drugs carry a very large increase in the risk of hip fracture. In fact, the study found that, in absolute terms, the increase in risk for regular users was small. Researchers found that among the women in the study who regularly used PPIs, about 2 in every 1,000 fractured a hip each year. In non-users, this figure was about 1.5 in every 1,000. This is a increase of about 5 fractures a year in every 10,000 women taking PPIs.

The Mail did point out this “absolute difference” towards the end of its story. Both the Mail and the BBC included comments from independent experts.

What kind of research was this?

The researchers point out that PPIs are among the most commonly used drugs worldwide. In the US they are available over the counter for general sale, but in the UK are available on prescription, and at the discretion of a pharmacist in certain situations without a prescription. They are used for symptoms of heartburn, gastro-oesophageal reflux disease (GORD) and stomach ulcers. PPIs are thought to work by reducing acid production in the stomach. Concern has grown over a potential association between long-term use of these drugs and bone fractures, although the researchers say that previous studies have had conflicting results and many did not take other factors (called confounders) that might affect the risk of fracture into account.

In their cohort study of nearly 80,000 post-menopausal women, the researchers set out to examine the association between long-term use of PPIs and the risk of hip fracture. Unlike a randomised controlled trial, a cohort study cannot prove cause and effect. However, cohort studies enable researchers to follow large groups of people for long periods and they are useful for looking at potential long-term risks and benefits of treatments. The study was prospective, which means it followed participants in time, rather than collecting information retrospectively. This makes it more reliable.

What did the research involve?

This study took its data from a large ongoing US study called the Nurses Health Study, which began in 1976 and which sent health questionnaires every two years to 121,700 female nurses aged 30-55.

From 1982 participants were asked to report all previous hip fractures and in each biennial questionnaire, women were asked if they had sustained a hip fracture over the previous two years. Those who reported a hip fracture were sent a follow-up questionnaire asking for more details. Fractures from bad accidents, such as falling down a flight of stairs, were excluded from the study. A review of medical records for 30 of the women validated all self-reported fractures.

From 2000 to 2006 the women were asked if they had regularly used a PPI in the previous two years. In earlier questionnaires (1994, 1996, 1998 and 2000), the women were also asked if they had regularly used other drugs for acid reflux, called H2 blockers.

The biennial questionnaires also included questions on other factors including menopausal status, body weight, leisure activities, smoking and alcohol use, use of hormone replacement therapy (HRT) and other medicines. Researchers used a validated food frequency questionnaire to calculate the women’s total intake of calcium and vitamin D.

They then analysed the data for any association between regular use of PPIs and hip fracture, adjusting their findings for key confounders such as body weight, physical activity, smoking, and alcohol and calcium intake. They also took into account whether the reasons for using a PPI might have affected the results.

Finally, they carried out a systematic review combining their results with 10 previous studies on the risk of hip fracture and the long-term use of PPIs.

What were the basic results?

The researchers documented 893 hip fractures during the period of the study. They also found that, in 2000, 6.7% of women regularly used a PPI – a figure that had risen to 18.9% by 2008.

  • Amongst women who had regularly taken a PPI at any time, there were 2.02 hip fractures per 1,000 person years, compared with 1.51 fractures per 1,000 person years among non-users.
  • Women who regularly used PPIs for at least two years had a 35% higher risk of hip fracture than non-users (age adjusted hazard ratio (HR) 1.35; 95% confidence interval (CI) 1.13 to 1.62), with longer use associated with increasing risk. Adjustment for risk factors, including body mass index, physical activity and intake of calcium did not alter this association (HR 1.36; CI 1.13 to 1.63).

The increased risk did not change when researchers also took into account the reasons for PPI use:

  • Current and former smokers who regularly used PPIs were 51% more likely to have a hip fracture than non-users (HR 1.51; (CI) 1.20 to 1.91).
  • Among women who never smoked there was no association between PPI use and hip fracture (HR 1.06; (CI) 0.77 to 1.46).
  • In a meta-analysis of these results with 10 previous studies, the risk of hip fracture in users of PPI was higher compared with non-users of PPIs (pooled odds ratio 1.30; CI 1.25 to 1.36).

The researchers also found that two years after women stopped taking PPIs, their risk of hip fracture returned to a similar level to that in women who had never taken them. Also, women taking H2 blockers had a “modest” increased risk of hip fracture but the risk was higher in women who took PPIs.

How did the researchers interpret the results?

The researchers conclude that their results provide “compelling evidence” of a risk between PPI use and hip fracture. They say the findings suggest that the need for long-term, continuous use of PPIs should be carefully evaluated, particularly among people who have smoked or are still smokers.

They suggest that PPIs may increase the risk of fracture by impairing the absorption of calcium, although in this study the risk of fracture was not affected by dietary calcium intake. The finding that the risk was confined to women with a history of smoking (an established risk factor for fracture) indicates that smoking and PPIs may act together (have a “synergistic effect”) on fracture risk.


This large study had several strengths. Unlike some previous studies, it collected information on and took into account other key risk factors for fracture, including body weight, smoking, alcohol use and physical activity. It also looked at the women’s use of PPIs every two years (rather than just asking them once) and took into account variations in use during this time in their analysis.

However, as the authors note, it also had some limitations:

  • It did not ask about the brands of PPI used, nor the doses of PPI the women took, both of which could affect risk of fracture.
  • The information about hip fracture was self-reported and not confirmed by medical records (although a smaller study has found self-reporting of hip fracture to be reliable).
  • Also, the study did not record the women’s bone mineral density (BMD). Low BMD is an important risk factor for fracture and adding a measure of this could have strengthened the study.

Finally, because this was a cohort study, other factors both measured and unmeasured may have affected the results, even though researchers took many of these into account in their analysis. Socio-economic status and education, for example, were not established. Because this was a study of registered nurses, the applicability of the results to other socio-economic groups might be limited.

This study found that the long-term, regular use of these drugs is associated with a small increased risk in hip fracture among older women, a risk that seems to be confined to past or current smokers. Women who regularly take PPIs and who are concerned about these findings are advised to talk to their GP or pharmacist. Further research will be needed to determine whether there is a need to revise the way these drugs are used.

NHS Attribution