To the Editor: Bolland and colleagues (1) fail to comment on the main issue I was addressing in the editorial, namely the discrepancy between the preoccupation with increasing a journal’s impact factor at the expense of a potentially harmful impact on patients (2). They demonstrate a breathtaking scientific arrogance by quoting themselves four times and not referring to other studies which were the focus of the letters of criticism and comment posted on the BMJ website. They incorrectly state ‘calcium supplements are commonly taken by older people for skeletal health’ (3) whereas the overwhelming majority take a combination with vitamin D which they did not study and which has not been associated with increased cardiovascular risk (4-7). As Nordin and colleagues point out (‘Making too much of a weak case’) Bolland’s study was in their view ‘seriously flawed’ and did not conform to the accepted criteria for meta-analyses, namely using only published data (8). In addition, no mention was made of the article by Lewis and colleagues where no increased risk was demonstrated (9). In fairness, in the Bolland article they do point out a substantial weakness in that none of the trials had cardiovascular outcomes as primary endpoints and cardiovascular data was not gathered in a standardised manner (3) – so it is rather difficult to draw any conclusions, as the analysis must therefore be incomplete. Though they report a significant increase in myocardial infarction (MI) over a median follow-up of only 3.6 years, which would imply remarkable previously unrecognised toxicity, there were no significant increases in the incidence of stroke or the composite end-point of MI, stroke or sudden death. The absence of an increase in death in spite of an apparent excess of MIs casts significant doubts on the validity of the observations. There also needs to be an explanation as to why the MI curves appear to part early at 1 year follow-up (figure 2) (3). This finding usually means unintentional risk discrepancy at baseline – they were not cardiac risk stratified. Even if valid, the relevance of their findings concerning calcium supplementation alone is rightly questioned because it is no longer recommended whereas a healthy diet is. So their study was really much ado about nothing, but for patients it translated into needless anxiety. I seem to remember a basic medical philosophy of first doing no harm. I do, however, completely agree with Bolland and colleagues when they state ‘we are concerned that, in some areas of medicine, too much emphasis is placed on analysis and reanalysis of a limited amount of trial level data, rather than on the design and conduct of informative randomised controlled trials in relevant populations’ (10). They go on to say the ‘conflicting conclusions thus generated can cause considerable confusion among health practitioners’ without mentioning patients, who are rendered both confused and fearful. The reason I do not reside in an armchair is because I spend my time turning people out of theirs. Perhaps, if Bolland and colleagues had followed their own 2009 advice, this time it would not have been necessary. None.