Bone drugs and cancer rates

Monday September 6 2010

Using osteoporosis drugs for long periods may double the risk of oesophagus cancer, The Guardian reported.

The news story is based on a study examining the rates of various cancers in people using bisphosphonates, a family of drugs primarily used to treat osteoporosis. The researchers estimate that taking bisphosphonates for five years would increase the number of cases of oesophageal (throat) cancer from 1 in 1,000 people to 2 in 1,000. Rates of stomach or colorectal cancer were not found to be affected.

The study benefits from its large size, use of reliable records and the fact that it accounted for the influence of smoking and alcohol, both well-established risk factors for oesophageal cancer.

However, this research did not conclude that patients should stop taking bisphosphonates, and cancer rates remained low in people using the drugs. The risk of all drugs needs to be balanced against their benefits in the individual, and anyone with concerns about potential side effects of bisphosphonates should talk to their GP or pharmacist.

Where did the story come from?

The study was carried out by researchers from Oxford University and funded by the Medical Research Council and Cancer Research UK. It was published in the peer-reviewed British Medical Journal.

The research was very well covered by BBC News, The Guardian and The Daily Telegraph. All of these news sources highlighted that the absolute risk of developing throat cancer is relatively low, even among people taking bisphosphonate drugs. Additionally, The Guardian featured a quote from an MHRA (Medicines and Healthcare products Regulatory Agency) spokesperson, who said: “The evidence from the study was not strong enough to suggest a definite causal association between oral bisphosphonates and oesophageal cancer. However, in order to reduce the risk of oesophageal irritation it is important to carefully follow the instructions [for bisphosphonates].” The Daily Telegraph also reported that the MHRA said that there was no need for patients to stop taking bisphosphonate medicines on the basis of this study alone.

What kind of research was this?

This was a case-control study that looked at whether a risk of oesophageal (throat) cancer is associated with taking oral bisphosphonate drugs to treat osteoporosis. This study was nested, meaning that both the people affected by a particular condition (cases) and those who are unaffected (controls) are drawn from the same population.

The researchers say that a potential side effect of taking oral bisphosphonates for osteoporosis is inflammation of the throat and, in some people, ulcers in the throat. They say that recent case reports suggested a possible increase in the risk of oesophageal cancer in people taking these osteoporosis drugs. The researchers wanted to look at the risk in a larger group of people to see if this was truly the case or if results presented in case reports had occurred by chance.

What did the research involve?

The researchers used data from the General Practice Research Database, which contains anonymous patient records from around 6 million people in the UK. All GP consultations, test results, diagnoses, admissions to hospital and prescriptions are recorded. If a patient dies, the cause of death is also recorded.

The researchers searched for patients who were over 40 and had been diagnosed with either oesophageal cancer, stomach cancer or colorectal cancer. For each patient who had one of these cancers (the cases), the researchers selected five control subjects who had no record of these types of cancer, were of a similar age, recruited from the same area of the UK and who had been followed by the database for a similar period.

They looked through the prescription records and recorded the patients who had received at least one prescription of oral bisphosphonate for osteoporosis. They did not include patients who had been prescribed bisphosphonates for any other condition. They assessed how long people had been taking the drugs by looking at the time interval between the first and last prescriptions.

What were the basic results?

The researchers identified 2,954 men and women with oesophageal cancer, 2,018 with stomach cancer and 10,641 with colorectal cancer. The mean observation period was 7.5 years. The average age at the time of diagnosis was 72. Those with oesophageal and stomach cancer were more likely to be smokers than their matched controls. There were similar proportions of smokers in the control group and colorectal cancer patients.

Both the case group and control group had similar proportions of subjects who had used bisphosphonates, with approximately 3% of each group having been issued at least one prescription for oral bisphosphonates during the study period. People who were prescribed bisphosphonates were more likely to be older and female.

The researchers found that before the year 2000, most patients were prescribed a bisphosphonate called etidronate. In 2000, alendronate, a bisphosphonate which could be taken weekly, was introduced. By 2005, most patients receiving bisphosphonates either took the weekly alendronate or another weekly bisphosphonate called risedronate.

The researchers found that after they had adjusted for factors that could influence the risk of oesophageal cancer, such as smoking or alcohol consumption and a high BMI, oesophageal cancer risk was 30% greater in those who had been issued a prescription of bisphosphonate medication (relative risk [RR] 1.30, 95% confidence interval [CI] 1.02 to 1.66).

People who had been issued more than 10 prescriptions of oral bisphosphonates had a 93% increased risk of throat cancer compared with people who had never had a prescription of these drugs (RR 1.93, 95% CI 1.37 to 2.70).

Patients who had been taking oral bisphosphonates for over three years had more than double the risk of developing oesophageal cancer than people who had never taken these drugs (RR 2.24, 95% CI 1.47 to 3.43).

There was no increased risk of oesophageal cancer for patients who had taken the drugs for shorter periods of time or who had received fewer prescriptions. Separately analysing each type of bisphosphonate prescribed did not seem to reveal any differences in cancer risk, although it should be noted that the number of patients in each drug group may have been too small to allow for a reliable analysis (for example, only 17 cases had been prescribed risedronate).

The researchers found no association between bisphosphonate use and stomach or colorectal cancer.

How did the researchers interpret the results?

The researchers said that there was an increased risk of oesophageal cancer associated with bisphosphonates, but that this increased risk was “largely restricted to those with 10 or more prescriptions spanning many years”.


This was a large cohort study that looked at whether using oral bisphosphonates for osteoporosis increased the risk of oesophageal, stomach or colorectal cancer. The study was well conducted and accounted for other factors that could affect the risk of these types of cancer, most importantly smoking and alcohol consumption.

Although the study found that people taking many prescriptions of these drugs over a long time had an increased risk compared to patients who had never taken these drugs, there are limitations to this study, some of which were raised by the researchers. Points for consideration include:

  • The researchers had prescription data but no information on how the patients took the drugs. For example, this type of drug will contain safety information designed to minimise irritation to the throat and it is not known whether this advice was followed. Also, it is not known how closely the patients followed the dosage and frequency instructions provided with their medication.
  • The researchers did not have information on whether patients had received bisphosphonates before they were included in the database.
  • Overall, 90 people with oesophageal cancer and 345 controls had previously used bisphosphonates. However, when dividing these people into subgroups based on the number of past prescriptions and duration of use, the sample sizes become smaller and, therefore, less statistically reliable. For example, although a more-than-double increased risk of cancer was found for three or more years of use, only 33 cases and 76 controls had used bisphosphonates for this period. There is an increased risk of finding false associations when calculating risk differences based on such small numbers of people.
  • The study only quoted relative risk increases. The absolute risks of developing throat cancer were not detailed in the UK population receiving bisphosphonates. However, the researchers used European and American cancer figures from the World Health Organization to estimate that bisphosphonate use would be associated with a doubling of cases of throat cancer in people aged 60-79, from 1 case in 1,000 people to 2 cases in 1,000.

The Guardian and The Daily Telegraph quoted the MHRA advice that there is no need to stop taking bisphosphonates based on this study alone. However, they say that throat irritation is a side effect of these drugs and that patients should carefully follow the safety instructions to minimise the risk of this side effect.

Analysis by Bazian
Edited by NHS Choices