Calcium supplements and cardiovascular disease - picking the spin

Abstract

To the Editor: We were roused from our armchairs by Jackson’s wearily sardonic dismissal of our work on calcium supplements and cardiovascular events (1,2). His editorial highlights the difficulties that surround the investigation of adverse effects of interventions (3,4). We invite his views on how he would have proceeded in the circumstances we encountered. We prespecified cardiovascular events as secondary end-points in a 5-year randomised controlled trial of calcium supplements in older women. The results of this analysis, which involved a comprehensive search for, and independent adjudication of all incident events, suggested that calcium supplements might increase the risk of myocardial infarction (MI) by 49% and of stroke by 37% (5). In a second smaller 2-year randomised controlled trial of calcium supplements in 323 men, all nine cardiac adverse events occurred in those allocated to calcium (6). The results were of marginal statistical significance, but the trials were underpowered. Would Jackson therefore have dismissed them and retired to his own armchair? We thought it quite important to investigate further, as calcium supplements are very widely prescribed and their efficacy in preventing fractures is sufficiently marginal (7) that any benefit accruing from their use would be countered by an increase in risk of MI of the magnitude suggested by our trials. There were no ongoing trials of calcium supplements that could address the issue and the logistical, ethical and funding problems associated with establishing a trial to investigate the hypothesis that calcium supplements cause harm are formidable. The best available option was to conduct the most extensive analysis possible of data available from completed clinical trials. This analysis also provided evidence that use of calcium supplements increases the risk of MI, a finding that was very consistent across the contributing trials (2). Of course, there are limitations to this work (acknowledged in the publication), as there are with any piece of clinical research. Would Jackson have sunk deeper in the armchair, turned up the music and dismissed these results also? Should, as Jackson implies, we only be concerned about interventions that are shown to increase mortality? Perhaps Jackson’s views were influenced by the body of evidence for efficacy of calcium supplements? In pooled analyses of trials of calcium supplements, the majority of which contributed data to our meta-analysis of cardiovascular events, the relative risk (95% CI) for fracture was 0.90 (0.80–1.00, nine studies, 6,517 participants). The addition of vitamin D makes little difference: relative risk of fracture 0.87 (0.77–0.97, eight studies, 46,108 participants) (7). These results hardly provide a compelling rationale for widespread prescription of calcium supplements, in the context of evidence of vascular harm. We should remember that both sides have spin bowlers. When facing a spin bowler, a cursory assessment precludes picking the direction of rotation, potentially leaving you stumped. None.