Osteoporosis is a serious condition in which the bones become fragile. It affects about 3 million people in the UK; most common in post-menopausal women, but younger women and men can be affected as well.
Superficially, bones look inert, but in fact they are constantly “turning over”: existing bone is broken down and new bone is formed. That’s why it is important for everyone, not just growing children, to have enough calcium in their diet, as calcium is a major constituent of bone. If new bone is not being formed as rapidly as old bone is breaking down, then the bone density will decrease and the bone will be weaker and more likely to break.
There are a number of drugs that can be used to help maintain bone density, and among the most common are a group called bisphosphonates. Proctor & Gamble (P&G) sells one called Actonel (generic name risedondrate) and its main rival is Novartis’ Fosamax (generic name alendronate).
Of the two, Fosamax is generally understood to be the more effective at reducing turnover and increasing bone density. Despite this, P&G hoped to be able to show that their drug was still as effective as Fosamax at reducing the risk of fractures, which is what matters. The risk of breaking a bone depends on a number of factors, not just the rate of calcium turnover, and it might well be that beyond a certain point, reducing turnover has no effect.
So P&G embarked on a large clinical trial, collaborating with researchers at Sheffield University in the UK. The results were published in a leading peer reviewed journal in the field [1]. The authors claim that there is indeed a threshold, and they provide graphs that appear to show this. They conclude that there is a level of bone turnover reduction beyond which no further fracture benefit is observed, just as P&G had hoped.
There isn’t enough information in the paper for one to repeat the calculation and construct the graphs, but then you wouldn’t necessarily expect there to be. However, when the lead author, Professor Richard Eastell of Sheffield University, presented the data at a meeting of the International Osteoporosis Foundation in 2002, an American investigator did ask how Eastell had arrived at his conclusion and was surprised to be told that he did not know, as all the analysis of the data had been carried out by P&G’s statistician.
According to the report in the Times Higher Educational Supplement (THES), Professor Eastell then suggested that in future the Sheffield group should carry out the analyses in parallel with P&G. P&G refused, on the curious grounds that while it “might add an extra layer of credibility,” it would also mean that “industry loses the opportunity to demonstrate its ability to be a true partner in scientific endeavours.” P&G also told the THES that it was standard industry practice to limit access to raw drug trial data [2].
That might have ended the matter, but the Sheffield group had already contracted with P&G to carry out a further study with Dr Aubrey Blumsohn taking the lead. When the measurements had been completed, Bluhmsohn began to feel uncomfortable because without the raw data he had no intuition for what was going on. (These were Phase 3 trials (see Box 1) and so were double blind. This means that the people carrying out the investigation could not know the results unless they were given the key, which P&G held.) Eastell asked P&G to allow Blumsohn access to the data, but this was refused. Despite this, three abstracts were prepared by P&G with Blumsohn listed as lead author.
Eventually, Blumsohn was allowed to visit the P&G laboratories to discuss the data. When he saw them, he began to have serious doubts about the conclusions. He complained that about 40 per cent of the data set was not displayed on the graphs, and he said that he was now not at all convinced that there really was a threshold, as the paper claimed.
Blumsohn and P&G were unable to come to an agreement on a way forward. Eventually he raised his concerns with Sheffield University. Dissatisfied with the university’s response, he then went to the Times Higher Educational Supplement, which has since published several articles on the affair [3]. The University has responded by instituting disciplinary proceedings against him for not following its internal procedures.
It is not obvious why Blumsohn should have gone through the internal procedures of the university, because the issue concerns an outside body. He himself says that he approached the university in the first place only to find out if it would support him if he were to challenge P&G.
Besides, universities in general are not known for standing up to pressure from companies that have a lot of money to spend. Think of the long struggle that Nancy Olivieri had to keep her job at the University of Toronto when her research proved embarrassing for Apotex, the manufacturer of the drug she was studying [4].
Since the reports appeared in the media, things have started to happen. P&G has issued what they call a “bill of rights” for researchers [5, 6]. Its academic collaborators will now have access to all the data relevant to their work with P&G, they will have final authority over all publication content, and company sponsored ghostwriters will provide help with writing papers “only if requested.”
That’s clearly a step forward, but it should have been standard practice all along, not something that had to be dragged out of P&G by pressure from the media. We also do not know how the bill of rights will work in practice. Blumsohn’s lawyers say that P&G is still withholding critical data from his study. For their part, P&G claim there is no connection between Blumsohn’s allegations and their new bill of rights [6].
In the meantime, think what all this tells us about drugs that are already on the market. If P&G’s behaviour was typical of the pharmaceutical industry, and they insist that it was, there must be many papers reporting the results of clinical trials that have appeared in peer reviewed journals and purport to be collaborations between industry and universities, often with the academics as lead authors (and therefore, presumably, as lead investigators) in which the academics have never even seen the raw data, still less been allowed to analyse it. The papers were written by ghost writers, hired by the company, who were not involved in the research and who are not listed as authors.
How confident can we now be that the drugs are as effective as the manufacturers claim?
Article first published 26/05/06
Got something to say about this page? Comment