There has been a 20-fold rise in the prescribing of bisphosphonates in general and alendronate in particular in recent years. In 1992 0·2% of women over 40 included in the UK General Practice Research Database (GPRD) were prescribed a bisphosphonate but by 2005 this had risen to 4·1%. At the same time there has been a parallel reduction in hormone replacement therapy (HRT) prescribing from 8·2% in 1991 to 7·0% in 2005, with a 50% fall since 2002, largely driven by concerns over excess risk of breast cancer (and to a lesser degree ovarian cancer) and cardiovascular events.
Bisphosphonates and particularly alendronate are well known to cause both dyspepsia and inflammatory changes such as erosive oesophagitis, delayed healing, and mucosal abnormalities.
Multinucleated giant cells have been detected in oesophageal inflammatory exudates. Whether these can undergo malignant transformation is not known but concerns have been raised as to a possible link between bisphosphonate use and upper GI cancer.
, in her report to the US Federal Drug Administration (FDA), noted that since the initial marketing of alendronate in 1995 the FDA had received 23 reports of patients who developed oesophageal tumours after taking the drug. Typically 2 years elapsed between the time patients started taking the drug and the onset of oesophageal cancer. In Europe and Japan a further 31 cases had been reported linking oesophageal cancer and bisphosphonate use. The FDA reporting did not include any denominator data. Wysowski suggested distal oesophageal carcinoma might be associated with bisphosphonate use but recommended more rigorous study approaches with sufficient size, length of follow up, inclusion of a control group, and control for confounding variables. In the meantime the authors advised that the drugs should not be used in patients with Barrett's oesophagus (an abnormal change in the cells of the lower portion of the oesophagus, thought to be due mainly to chronic acid reflux from the stomach, its main significance is an increased risk of developing adenocarcinoma of the oesophagus). However, Merck (the makers of alendronate) have not reported any cases of oesophageal cancer linked to bisphosphonate use in their clinical database of 17,000 patients.
Following the FDA reports, rapid communications of studies using a large national database in Denmark and Medicare beneficiaries in the US concluded there was no evidence for an increased risk of oesophageal cancer in bisphosphonate users
. However, again the follow up period was short (2 years). Because of the relative rarity of these conditions and the limited number of patients, the confidence intervals were wide and it was impossible to conclude whether there could be a clinically important association. Two published studies using the UK General Practice Research Database (GPRD) reached differing conclusions. Cardwell et al.
in a retrospective cohort study found no evidence of an increase in the combined risk for gastric and oesophageal cancer in bisphosphonate users compared to non-users, Hazard Ratio 0·96, (95% CI 0·74–1·49) but again confidence intervals were wide (i.e. study is also consistent with a fairly substantial effect). Green et al.
carried out a nested case control study using a sample drawn from the same database and found an overall increased risk of oesophageal cancer in bisphosphonate users (Relative Risk 1·30, 95% CI 1·02–1·66). This increased with more than ten prescriptions or longer than three years use (RR 1·93). Although these studies appear to give different results, the relatively wide confidence intervals overlap substantially, so results could be consistent with a similar magnitude of risk. In Cardwell's study a relatively small proportion of exposed cases were included limiting power and precision.
Dixon and Solomon 
have reviewed in detail the conflicting results of these two studies and concluded similarly that even when confining the comparison to patients with greater than three years exposure to bisphosphonates (allowing for induction and latency periods) the confidence intervals of the relative risks for both studies included the possibility of a 50% increase in risk. At less than three years exposure, although neither study found an increase in relative risk of oesophageal cancer with bisphosphonate use, Dixon and Solomon point out that the confidence intervals of both span one and, therefore, one should say that the results are ‘inconclusive’ rather than “there is no effect”. The upper limits of the RR's of 3 years' exposure are 1.73 and 1.81 respectively according to the 95% confidence intervals of the two analyses. They note “it is, therefore, plausible in both studies that, despite the best guess being of ‘no increased risk’, there may be as much as a 70% increase in baseline risk.”
Haber et al
concluded after reviewing both Green and Cardwell's studies and the observational study on the incidence of oesophageal cancer in patients with Barrett's oesophagus taking bisphosphonates by Nugyen
, as well as the case reports in Wysowski's article, that “the evidence on the use of bisphosphonates and risk of esophageal cancer is weak and conflicting”.
Our primary aim was to carry out a retrospective case control study with the greatest possible power using a large UK primary care database (the GPRD), to determine whether any association exists between prescribing of alendronate specifically (and bisphosphonates in general) and the development of upper GI malignancy. We were particularly concerned to have enough statistical power to detect a small increase in a rare, but serious disease and, therefore, chose to do a matched case control study using all known cases of oesophageal and gastric cancer. Although this is the third study using data drawn from the GRPD, it uses a later and significantly larger version of the database and it is substantially more powerful.