20 February 2012
Evidence from a large cohort study suggests that postmenopausal women who were previous or current smokers and who regularly used PPIs for more than two years had about a 50% increased relative risk of low-trauma hip fracture. However, in absolute terms the risk was small: about 8 extra fractures per 10,000 women years. This was independent of risk factors such as dietary calcium intake, physical activity, history of osteoporosis or other drugs known to affect the risk of hip fracture. Women who had stopped using PPIs more than two years previously had a risk of fracture similar to that of women who had never used a PPI.
Level of evidence:
Level 2 (limited quality patient-oriented evidence) according to the SORT criteria.
Health professionals should carefully review PPI prescribing and make sure it is in line with the relevant NICE guidance. PPIs have several important places in therapy. Nevertheless, PPIs should not be initiated or continued long term without careful thought and a discussion with the patient of the likely risks and benefits, including these new data around hip fracture.
What is the background to this?
The FDA announced in June 2010 that it was revising product information for PPIs to include new safety information about a possible increased risk of fractures of the hip, wrist, and spine. This was based on the FDA’s review of several epidemiological studies. Some studies found that those at greatest risk for these fractures received high doses of PPIs or used them for one year or more (see the FDA’s data summary). The majority of the studies evaluated individuals 50 years of age or older and the increased risk of fracture was primarily observed in this age group. Randomised controlled trials (RCTs) of PPIs have not shown an increased risk of fractures but these trials have generally lasted only 6 months.
Many of the epidemiological studies had important limitations, including an inability to control for important dietary and lifestyle confounders, small sample size, and limited ascertainment of PPI exposure. The authors of the study discussed here therefore sought to examine the association between long term PPI use and risk of hip fracture among postmenopausal women enrolled in a large prospective cohort, the Nurses’ Health Study, in which detailed information about dietary and lifestyle factors are collected every two years.
What does this study claim?
After excluding women who had previously reported hip fracture or cancer, or hip fracture associated with high trauma, the absolute risk of hip fracture among regular users of PPIs was 2.02 events per 1000 women years, compared with 1.51 events per 1000 women years among non-users. Compared with non-users, regular users of PPIs had a 35% increase in risk of hip fracture (age–adjusted hazard ratio [HR] 1.35, 95% confidence interval [CI] 1.13 to 1.62). This increased risk remained largely unchanged in a multivariate analysis which adjusted for age, body mass index (BMI), alcohol intake, total energy adjusted calcium intake, history of osteoporosis, level of physical activity, smoking status, vitamin D intake, or use of bisphosphonates, thiazides, corticosteroids, and hormone replacement therapy (HRT): HR 1.36, 95% CI 1.13 to 1.63.
Perhaps surprisingly, when the contributing effect of several known risk factors (age, BMI, use of HRT, use of corticosteroids, history of osteoporosis, intake of calcium and smoking history) were examined, the risk of hip fracture in PPI-users differed from non-users to a statistically significant extent only according to history of cigarette smoking. In a model which adjusted for all the other risk factors in the paragraph above, women who either previously or currently smoked had an HR for fracture of 1.51 (95%CI 1.20 to 1.91). In contrast, there was no significant association between PPI use and risk of fracture among women who never smoked (HR 1.06, 95%CI 0.77 to 1.46, p for interaction 0.03). Applying the HR for ‘ever-smokers’ to the baseline risk of PPI non-users suggests an absolute increase of 0.77 (95% CI 0.3 to 1.37) fractures per 1000 women years among such women.
The authors also considered the time since stopping PPI use in relation to risk of fracture. Although risk persisted for women who stopped PPI use within 2 years, women who had stopped use more than 2 years previously had a fully adjusted HR of 1.10 (95% CI 0.63 to 1.92), compared to women who had never used a PPI.
These data add to the evidence about the risks of PPIs. The size of this study, its methodology, duration and the number of events recorded are impressive. However, observational studies such as this one can only prove association not establish causation, because they are prone to confounding. Unlike in the setting of an RCT, in ‘real life’ treatment plans are chosen, changed, or actively not chosen in the light of individual patients’ risk factors, preferences and tolerability or response to other drugs. Thus observed differences in outcomes may be due to differences among the patients, not only the different treatments.
In this study, considerable efforts were made to reduce confounding and adjust for known risk factors and other variables. However, by definition, the authors could not adjust for unknown risk factors. However, the authors conducted a meta-analysis of their study with 10 other epidemiological studies and found a similar overall hazard ratio. The study’s finding that the risk of fracture appears to be confined to women with a history of smoking, an established risk factor for fracture, is biologically plausible (smoking and PPI use may both affect calcium absorption, with a potentially synergistic effect). It is clearly not possible to know from this study if men who use PPIs have the same, higher or lower risk of hip fracture than male non-users.
When discussing this study with women who are concerned about their risk, it is important to point out that this appears, from this study, to be confined to postmenopausal women ‘ever-smokers’ who used PPIs for more than 2 years. Health professionals should also point out that the absolute risk increase was small (about 8 extra hip fractures per 10,000 women years).
However, even modest increases in risks of adverse events add up to a substantial amount of patient harm at a population level when the risk factor (e.g. PPI exposure) is widely experienced. More than 10 million prescriptions for PPIs were dispensed in England during the last quarter for which data are available (January to March 2011: Personal communication, NHSBSA PPD).
In addition to possible increased risks of hip fracture, there are other risks associated with PPI use. A large observational study found that hospital inpatients taking daily PPIs were over 70% more likely to develop Clostridium difficile infection than non-users. A second US study found that people who already have C difficile infection and are treated with PPIs had a more than 40% increased relative risk of recurrence of infection. These are discussed in MeReC Rapid Review 1494. More recently, the FDA has warned that a diagnosis of C difficile-associated diarrhoea should be considered for patients taking PPIs who develop diarrhoea that does not improve.
However, it is important to set these risks against the clear benefits that result from PPI treatment in certain indications. NICE guidance on dyspepsia recommends that patients presenting in primary care with dyspepsia in whom urgent referral for endoscopy is not indicated should initially have their medication reviewed and be offered lifestyle advice. If this is not successful, they should be offered either one month of full dose PPI, or testing for H pylori infection and eradication treatment. If the option tried is not successful, the alternative should be offered. If patients still experience symptoms, NICE advises health professionals to offer low-dose treatment with a limited number of repeat prescriptions. They should discuss the use of treatment on an as-required basis to help patients manage their own symptoms. Similarly, after initial treatment, patients with endoscopically-determined GORD or peptic ulceration should continue on low dose treatment, as required, only if necessary and with a limited number of repeat prescriptions. Patients should be reviewed at least annually. NICE guidance on low back pain, osteoarthritis and rheumatoid arthritis advise that in these conditions, NSAIDs should be co-prescribed with a PPI [if over 45 among those with low back pain] to reduce the risk of adverse gastrointestinal side effects.
Khalili H et al. Use of proton pump inhibitors and risk of hip fracture in relation to dietary and lifestyle factors: a prospective cohort study. BMJ2012;344doi: 10.1136/bmj.e372 (Published 31 January 2012)
Design Prospective cohort study, drawing data from the Nurses’ Health Study
Patients US female registered nurses Women in the study receive a detailed questionnaire on a wide range of health and lifestyle matters every two years, which is followed up if more information is required. Women have been asked about the use of PPIs since the questionnaire in 2000. After excluding women who had previously reported hip fracture or cancer, or hip fracture associated with high trauma (such as from skiing or falling down a flight of stairs), reports for 79,899 women covering the period 2000 to 2008 were available. These included 893 incident hip fractures over 565,786 person years of follow-up.
Outcomes and results See above
Study sponsorship The study was funded by the US National Institutes of Health.
Please comment on this rapid review using our feedback form.
Make sure you are signed up to NPC Email updates — the free email alerting system that keeps you up to date with the NPC news and outputs relevant to you.